Script for easily configuring, using, switching and comparing local offline coding models
Initial commit (before AI edits)
+2898
+22
.claude/settings.local.json
+22
.claude/settings.local.json
···
1
+
{
2
+
"permissions": {
3
+
"allow": [
4
+
"Bash(~/.local/bin/llama-chat-server:*)",
5
+
"Bash(curl:*)",
6
+
"Bash(llama-server:*)",
7
+
"Bash(pkill:*)",
8
+
"Bash(sysctl:*)",
9
+
"WebFetch(domain:github.com)",
10
+
"WebFetch(domain:raw.githubusercontent.com)",
11
+
"Bash(pipx install:*)",
12
+
"Bash(python3.12:*)",
13
+
"Bash(HOMEBREW_NO_AUTO_UPDATE=1 brew install:*)",
14
+
"Bash(go install:*)",
15
+
"Bash(~/.local/bin/localcode help:*)",
16
+
"Bash(~/.local/bin/localcode status:*)",
17
+
"Bash(~/.local/bin/localcode models:*)",
18
+
"Bash(~/.local/bin/localcode tuis:*)",
19
+
"Bash(~/.local/bin/localcode bench:*)"
20
+
]
21
+
}
22
+
}
+91
CLAUDE.md
+91
CLAUDE.md
···
1
+
# CLAUDE.md
2
+
3
+
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
4
+
5
+
## Project Overview
6
+
7
+
`localcode` — a single CLI for managing a fully offline local AI coding environment on macOS Apple Silicon. Uses llama.cpp to serve Qwen 2.5 Coder models via OpenAI-compatible APIs, with switchable terminal coding agents (Aider, OpenCode, Pi).
8
+
9
+
## Commands
10
+
11
+
- `npm run dev -- <args>` — Run via tsx (development)
12
+
- `npm run build` — Compile TypeScript to `dist/`
13
+
- `npx tsc --noEmit` — Type-check without emitting
14
+
15
+
After `localcode setup`, the `localcode` binary is available in `~/.local/bin/`.
16
+
17
+
## CLI
18
+
19
+
```
20
+
localcode Launch active TUI in current directory
21
+
localcode status Show current config + server health
22
+
localcode start Start chat + autocomplete servers
23
+
localcode stop Stop all servers
24
+
localcode models List available models
25
+
localcode models set-chat <id> Switch chat model
26
+
localcode models set-auto <id> Switch autocomplete model
27
+
localcode tuis List available TUIs
28
+
localcode tuis set <id> Switch active TUI
29
+
localcode bench Benchmark running chat model
30
+
localcode bench history Show past benchmark results
31
+
localcode pipe "prompt" Pipe stdin through the model
32
+
localcode setup Full install
33
+
```
34
+
35
+
## Architecture
36
+
37
+
```
38
+
src/
39
+
main.ts — CLI dispatcher (switch on process.argv[2])
40
+
config.ts — Path/port constants
41
+
log.ts — log/warn/err with ANSI colors
42
+
util.ts — Shell exec helpers, file writers
43
+
runtime-config.ts — Read/write ~/.config/localcode/config.json
44
+
registry/
45
+
models.ts — ModelDef interface + MODELS array
46
+
tuis.ts — TuiDef interface + TUIS array
47
+
commands/
48
+
run.ts — Default action: ensure server, init git, exec TUI
49
+
status.ts — Show config + server health
50
+
server.ts — Start/stop llama.cpp servers
51
+
setup.ts — Full install pipeline
52
+
models.ts — List/switch models, auto-download + regen scripts
53
+
tuis.ts — List/switch TUIs, auto-install + regen scripts
54
+
bench.ts — Benchmark against running llama.cpp server
55
+
pipe.ts — Pipe stdin through the model
56
+
steps/ — Individual setup phases (preflight, homebrew, llama, etc.)
57
+
templates/
58
+
scripts.ts — Bash server launcher templates (parameterized by ModelDef)
59
+
aider.ts — Aider config template
60
+
opencode.ts — OpenCode config template
61
+
pi.ts — Pi models.json template
62
+
```
63
+
64
+
### Key patterns
65
+
66
+
**Runtime config** (`~/.config/localcode/config.json`): Stores active chatModel, autocompleteModel, and tui IDs. Read by `runtime-config.ts` with defaults fallback.
67
+
68
+
**Registries**: `registry/models.ts` and `registry/tuis.ts` define available options as typed arrays. Add new models/TUIs by appending to these arrays.
69
+
70
+
**Script regeneration**: When models or TUI are switched, launcher scripts in `~/.local/bin/` and all TUI configs are automatically regenerated.
71
+
72
+
**Template escaping**: Bash templates in `src/templates/scripts.ts` use `const D = "$"` to emit literal `$` without triggering TS interpolation.
73
+
74
+
**Generated scripts**: Only 3 bash scripts are generated in `~/.local/bin/`: `localcode` (thin wrapper calling `node dist/main.js`), `llama-chat-server`, `llama-complete-server`. All other functionality lives in TypeScript commands.
75
+
76
+
**Benchmark**: Hits `/v1/chat/completions` with 3 hardcoded prompts, measures wall-clock time + token counts. Results saved to `~/.config/localcode/benchmarks.json`.
77
+
78
+
## Key paths on the user's system
79
+
80
+
- `~/.local/bin/` — `localcode` wrapper + server launcher scripts
81
+
- `~/.local/share/llama-models/` — Downloaded GGUF model files
82
+
- `~/.config/localcode/config.json` — Active model/TUI selection
83
+
- `~/.config/localcode/benchmarks.json` — Benchmark history
84
+
- `~/.aider/` — Aider config
85
+
- `~/.config/opencode/opencode.json` — OpenCode config
86
+
- `~/.pi/agent/models.json` — Pi config
87
+
- Chat server port **8080**, autocomplete port **8081**
88
+
89
+
## Important: after changing TypeScript
90
+
91
+
The `localcode` wrapper in `~/.local/bin/` calls `node dist/main.js`. After modifying TypeScript source, run `npm run build` to recompile, or the wrapper will run stale code.
+192
README.md
+192
README.md
···
1
+
# Local AI Coding Environment
2
+
3
+
A fully offline, privacy-first AI coding setup for macOS Apple Silicon. Uses **llama.cpp** to run **Qwen 2.5 Coder** models locally, with **Aider** and **OpenCode** as terminal-based coding agents — no API keys, no cloud, no costs.
4
+
5
+
## Hardware Requirements
6
+
7
+
- **Mac with Apple Silicon** (M1/M2/M3/M4)
8
+
- **32GB RAM** recommended (the 32B model uses ~20GB)
9
+
- ~25GB free disk space for models
10
+
11
+
## What Gets Installed
12
+
13
+
| Component | Purpose |
14
+
|---|---|
15
+
| **llama.cpp** | Local model inference with Metal GPU acceleration |
16
+
| **Qwen 2.5 Coder 32B** (Q4_K_M) | Main chat/coding model (~20GB) |
17
+
| **Qwen 2.5 Coder 1.5B** (Q4_K_M) | Fast autocomplete model (~1.2GB) |
18
+
| **Aider** | Terminal coding agent (Claude Code alternative) |
19
+
| **jq** | JSON processing for the pipe command |
20
+
21
+
Both models are served via llama.cpp's built-in OpenAI-compatible API, making them work with any tool that supports the OpenAI API format.
22
+
23
+
## Installation
24
+
25
+
```bash
26
+
chmod +x setup.sh
27
+
./setup.sh
28
+
```
29
+
30
+
The script is idempotent — safe to run multiple times. The first run downloads ~21GB of model weights from HuggingFace.
31
+
32
+
After installation, restart your shell or run:
33
+
34
+
```bash
35
+
source ~/.zshrc
36
+
```
37
+
38
+
## Commands
39
+
40
+
### `llama-start`
41
+
42
+
Starts both llama.cpp servers in the foreground. Press `Ctrl+C` to stop both.
43
+
44
+
- Chat model (32B): `http://127.0.0.1:8080`
45
+
- Autocomplete model (1.5B): `http://127.0.0.1:8081`
46
+
47
+
### `llama-stop`
48
+
49
+
Kills all running llama-server processes.
50
+
51
+
### `ai-code [directory]`
52
+
53
+
The main coding agent. Auto-starts the chat server if it's not running. Initializes a git repo if one doesn't exist, then launches Aider with full file-editing capabilities.
54
+
55
+
```bash
56
+
cd ~/projects/my-app
57
+
ai-code .
58
+
59
+
# or from anywhere
60
+
ai-code ~/projects/my-app
61
+
```
62
+
63
+
Inside Aider you can ask it to edit files, run commands, and refactor code across your project. Changes are auto-committed to git so you can always roll back.
64
+
65
+
### `ai-ask "question"`
66
+
67
+
Quick Q&A mode — no file editing, just chat. Useful for coding questions without modifying your project.
68
+
69
+
```bash
70
+
ai-ask "how do I handle errors in rust"
71
+
```
72
+
73
+
### `ai-pipe "prompt"`
74
+
75
+
Pipe code through the model via stdin. Useful for one-shot transforms in scripts.
76
+
77
+
```bash
78
+
cat main.py | ai-pipe "add type hints"
79
+
git diff | ai-pipe "write a commit message"
80
+
cat api.go | ai-pipe "find bugs"
81
+
```
82
+
83
+
## Using OpenCode Instead of Aider
84
+
85
+
[OpenCode](https://opencode.ai) is another terminal coding agent with a polished TUI. It connects to the same llama.cpp backend.
86
+
87
+
### Install OpenCode
88
+
89
+
```bash
90
+
brew install anomalyco/tap/opencode
91
+
```
92
+
93
+
### Configure
94
+
95
+
Create `~/.config/opencode/opencode.json`:
96
+
97
+
```json
98
+
{
99
+
"$schema": "https://opencode.ai/config.json",
100
+
"model": "llama-cpp/qwen2.5-coder-32b",
101
+
"provider": {
102
+
"llama-cpp": {
103
+
"npm": "@ai-sdk/openai-compatible",
104
+
"name": "llama.cpp (local)",
105
+
"options": {
106
+
"baseURL": "http://127.0.0.1:8080/v1",
107
+
"apiKey": "not-needed"
108
+
},
109
+
"models": {
110
+
"qwen2.5-coder-32b": {
111
+
"name": "Qwen 2.5 Coder 32B",
112
+
"tools": true
113
+
}
114
+
}
115
+
}
116
+
}
117
+
}
118
+
```
119
+
120
+
### Run
121
+
122
+
```bash
123
+
llama-start # start the server (or ai-code auto-starts it)
124
+
opencode # launch OpenCode in your project directory
125
+
```
126
+
127
+
Use `/models` inside OpenCode to select the Qwen model, and `Tab` to switch between Plan and Build modes.
128
+
129
+
## Configuration Files
130
+
131
+
| File | Purpose |
132
+
|---|---|
133
+
| `~/.aider/aider.conf.yml` | Aider settings (model, git, UI) |
134
+
| `~/.aider/.env` | API base URL and key for Aider |
135
+
| `~/.config/opencode/opencode.json` | OpenCode provider config |
136
+
| `~/.local/share/llama-models/` | Downloaded GGUF model files |
137
+
| `~/.local/bin/` | Launcher scripts |
138
+
139
+
## llama.cpp Server Flags
140
+
141
+
The chat server launches with these defaults:
142
+
143
+
| Flag | Value | Purpose |
144
+
|---|---|---|
145
+
| `--ctx-size` | 16384 | Context window (increase to 32768 if tools misbehave) |
146
+
| `--n-gpu-layers` | 99 | Offload all layers to Metal GPU |
147
+
| `--flash-attn` | — | Enable flash attention for speed |
148
+
| `--mlock` | — | Lock model in RAM to prevent swapping |
149
+
| `--threads` | auto | Uses performance core count |
150
+
151
+
## Troubleshooting
152
+
153
+
**Model loading is slow on first run**: The first inference after starting the server takes 10–30 seconds while the model loads into memory. Subsequent requests are fast.
154
+
155
+
**Running out of RAM / swapping**: The 32B Q4 model needs ~20GB. Close memory-heavy apps. You can also try the smaller `qwen2.5-coder-14b-instruct-q4_k_m.gguf` instead.
156
+
157
+
**OpenCode tools not working**: Increase `--ctx-size` to 32768 in the `llama-chat-server` script. Tool-calling needs more context to work reliably.
158
+
159
+
**Slow generation speed**: Expect ~15–25 tokens/sec on the 32B model with M4. This is normal for a model this size running locally. The 1.5B autocomplete model runs much faster.
160
+
161
+
**Server won't start**: Check if another process is using port 8080 or 8081 with `lsof -i :8080`. Use `llama-stop` to kill stale processes.
162
+
163
+
## Performance Expectations
164
+
165
+
| Model | Speed | Use Case |
166
+
|---|---|---|
167
+
| Qwen 2.5 Coder 32B | ~15–25 tok/s | Chat, code generation, refactoring |
168
+
| Qwen 2.5 Coder 1.5B | ~100+ tok/s | Autocomplete, quick suggestions |
169
+
170
+
Both models run entirely on-device using Metal acceleration. No network connection required after initial setup.
171
+
172
+
## Uninstall
173
+
174
+
```bash
175
+
# Remove models (~21GB)
176
+
rm -rf ~/.local/share/llama-models
177
+
178
+
# Remove launcher scripts
179
+
rm ~/.local/bin/{ai-code,ai-ask,ai-pipe,llama-start,llama-stop,llama-chat-server,llama-complete-server}
180
+
181
+
# Remove configs
182
+
rm -rf ~/.aider
183
+
rm -f ~/.config/opencode/opencode.json
184
+
185
+
# Remove Ollama auto-start (if set)
186
+
launchctl unload ~/Library/LaunchAgents/com.ollama.serve.plist
187
+
rm ~/Library/LaunchAgents/com.ollama.serve.plist
188
+
189
+
# Uninstall packages
190
+
pipx uninstall aider-chat
191
+
brew uninstall llama.cpp jq
192
+
```
+5
gen-scripts.ts
+5
gen-scripts.ts
+20
opencode-config.json
+20
opencode-config.json
···
1
+
{
2
+
"$schema": "https://opencode.ai/config.json",
3
+
"model": "llama-cpp/qwen2.5-coder-32b",
4
+
"provider": {
5
+
"llama-cpp": {
6
+
"npm": "@ai-sdk/openai-compatible",
7
+
"name": "llama.cpp (local)",
8
+
"options": {
9
+
"baseURL": "http://127.0.0.1:8080/v1",
10
+
"apiKey": "not-needed"
11
+
},
12
+
"models": {
13
+
"qwen2.5-coder-32b": {
14
+
"name": "Qwen 2.5 Coder 32B",
15
+
"tools": true
16
+
}
17
+
}
18
+
}
19
+
}
20
+
}
+590
package-lock.json
+590
package-lock.json
···
1
+
{
2
+
"name": "localcode-setup",
3
+
"version": "1.0.0",
4
+
"lockfileVersion": 3,
5
+
"requires": true,
6
+
"packages": {
7
+
"": {
8
+
"name": "localcode-setup",
9
+
"version": "1.0.0",
10
+
"devDependencies": {
11
+
"@types/node": "^22",
12
+
"tsx": "^4",
13
+
"typescript": "^5"
14
+
}
15
+
},
16
+
"node_modules/@esbuild/aix-ppc64": {
17
+
"version": "0.27.3",
18
+
"resolved": "https://registry.npmjs.org/@esbuild/aix-ppc64/-/aix-ppc64-0.27.3.tgz",
19
+
"integrity": "sha512-9fJMTNFTWZMh5qwrBItuziu834eOCUcEqymSH7pY+zoMVEZg3gcPuBNxH1EvfVYe9h0x/Ptw8KBzv7qxb7l8dg==",
20
+
"cpu": [
21
+
"ppc64"
22
+
],
23
+
"dev": true,
24
+
"license": "MIT",
25
+
"optional": true,
26
+
"os": [
27
+
"aix"
28
+
],
29
+
"engines": {
30
+
"node": ">=18"
31
+
}
32
+
},
33
+
"node_modules/@esbuild/android-arm": {
34
+
"version": "0.27.3",
35
+
"resolved": "https://registry.npmjs.org/@esbuild/android-arm/-/android-arm-0.27.3.tgz",
36
+
"integrity": "sha512-i5D1hPY7GIQmXlXhs2w8AWHhenb00+GxjxRncS2ZM7YNVGNfaMxgzSGuO8o8SJzRc/oZwU2bcScvVERk03QhzA==",
37
+
"cpu": [
38
+
"arm"
39
+
],
40
+
"dev": true,
41
+
"license": "MIT",
42
+
"optional": true,
43
+
"os": [
44
+
"android"
45
+
],
46
+
"engines": {
47
+
"node": ">=18"
48
+
}
49
+
},
50
+
"node_modules/@esbuild/android-arm64": {
51
+
"version": "0.27.3",
52
+
"resolved": "https://registry.npmjs.org/@esbuild/android-arm64/-/android-arm64-0.27.3.tgz",
53
+
"integrity": "sha512-YdghPYUmj/FX2SYKJ0OZxf+iaKgMsKHVPF1MAq/P8WirnSpCStzKJFjOjzsW0QQ7oIAiccHdcqjbHmJxRb/dmg==",
54
+
"cpu": [
55
+
"arm64"
56
+
],
57
+
"dev": true,
58
+
"license": "MIT",
59
+
"optional": true,
60
+
"os": [
61
+
"android"
62
+
],
63
+
"engines": {
64
+
"node": ">=18"
65
+
}
66
+
},
67
+
"node_modules/@esbuild/android-x64": {
68
+
"version": "0.27.3",
69
+
"resolved": "https://registry.npmjs.org/@esbuild/android-x64/-/android-x64-0.27.3.tgz",
70
+
"integrity": "sha512-IN/0BNTkHtk8lkOM8JWAYFg4ORxBkZQf9zXiEOfERX/CzxW3Vg1ewAhU7QSWQpVIzTW+b8Xy+lGzdYXV6UZObQ==",
71
+
"cpu": [
72
+
"x64"
73
+
],
74
+
"dev": true,
75
+
"license": "MIT",
76
+
"optional": true,
77
+
"os": [
78
+
"android"
79
+
],
80
+
"engines": {
81
+
"node": ">=18"
82
+
}
83
+
},
84
+
"node_modules/@esbuild/darwin-arm64": {
85
+
"version": "0.27.3",
86
+
"resolved": "https://registry.npmjs.org/@esbuild/darwin-arm64/-/darwin-arm64-0.27.3.tgz",
87
+
"integrity": "sha512-Re491k7ByTVRy0t3EKWajdLIr0gz2kKKfzafkth4Q8A5n1xTHrkqZgLLjFEHVD+AXdUGgQMq+Godfq45mGpCKg==",
88
+
"cpu": [
89
+
"arm64"
90
+
],
91
+
"dev": true,
92
+
"license": "MIT",
93
+
"optional": true,
94
+
"os": [
95
+
"darwin"
96
+
],
97
+
"engines": {
98
+
"node": ">=18"
99
+
}
100
+
},
101
+
"node_modules/@esbuild/darwin-x64": {
102
+
"version": "0.27.3",
103
+
"resolved": "https://registry.npmjs.org/@esbuild/darwin-x64/-/darwin-x64-0.27.3.tgz",
104
+
"integrity": "sha512-vHk/hA7/1AckjGzRqi6wbo+jaShzRowYip6rt6q7VYEDX4LEy1pZfDpdxCBnGtl+A5zq8iXDcyuxwtv3hNtHFg==",
105
+
"cpu": [
106
+
"x64"
107
+
],
108
+
"dev": true,
109
+
"license": "MIT",
110
+
"optional": true,
111
+
"os": [
112
+
"darwin"
113
+
],
114
+
"engines": {
115
+
"node": ">=18"
116
+
}
117
+
},
118
+
"node_modules/@esbuild/freebsd-arm64": {
119
+
"version": "0.27.3",
120
+
"resolved": "https://registry.npmjs.org/@esbuild/freebsd-arm64/-/freebsd-arm64-0.27.3.tgz",
121
+
"integrity": "sha512-ipTYM2fjt3kQAYOvo6vcxJx3nBYAzPjgTCk7QEgZG8AUO3ydUhvelmhrbOheMnGOlaSFUoHXB6un+A7q4ygY9w==",
122
+
"cpu": [
123
+
"arm64"
124
+
],
125
+
"dev": true,
126
+
"license": "MIT",
127
+
"optional": true,
128
+
"os": [
129
+
"freebsd"
130
+
],
131
+
"engines": {
132
+
"node": ">=18"
133
+
}
134
+
},
135
+
"node_modules/@esbuild/freebsd-x64": {
136
+
"version": "0.27.3",
137
+
"resolved": "https://registry.npmjs.org/@esbuild/freebsd-x64/-/freebsd-x64-0.27.3.tgz",
138
+
"integrity": "sha512-dDk0X87T7mI6U3K9VjWtHOXqwAMJBNN2r7bejDsc+j03SEjtD9HrOl8gVFByeM0aJksoUuUVU9TBaZa2rgj0oA==",
139
+
"cpu": [
140
+
"x64"
141
+
],
142
+
"dev": true,
143
+
"license": "MIT",
144
+
"optional": true,
145
+
"os": [
146
+
"freebsd"
147
+
],
148
+
"engines": {
149
+
"node": ">=18"
150
+
}
151
+
},
152
+
"node_modules/@esbuild/linux-arm": {
153
+
"version": "0.27.3",
154
+
"resolved": "https://registry.npmjs.org/@esbuild/linux-arm/-/linux-arm-0.27.3.tgz",
155
+
"integrity": "sha512-s6nPv2QkSupJwLYyfS+gwdirm0ukyTFNl3KTgZEAiJDd+iHZcbTPPcWCcRYH+WlNbwChgH2QkE9NSlNrMT8Gfw==",
156
+
"cpu": [
157
+
"arm"
158
+
],
159
+
"dev": true,
160
+
"license": "MIT",
161
+
"optional": true,
162
+
"os": [
163
+
"linux"
164
+
],
165
+
"engines": {
166
+
"node": ">=18"
167
+
}
168
+
},
169
+
"node_modules/@esbuild/linux-arm64": {
170
+
"version": "0.27.3",
171
+
"resolved": "https://registry.npmjs.org/@esbuild/linux-arm64/-/linux-arm64-0.27.3.tgz",
172
+
"integrity": "sha512-sZOuFz/xWnZ4KH3YfFrKCf1WyPZHakVzTiqji3WDc0BCl2kBwiJLCXpzLzUBLgmp4veFZdvN5ChW4Eq/8Fc2Fg==",
173
+
"cpu": [
174
+
"arm64"
175
+
],
176
+
"dev": true,
177
+
"license": "MIT",
178
+
"optional": true,
179
+
"os": [
180
+
"linux"
181
+
],
182
+
"engines": {
183
+
"node": ">=18"
184
+
}
185
+
},
186
+
"node_modules/@esbuild/linux-ia32": {
187
+
"version": "0.27.3",
188
+
"resolved": "https://registry.npmjs.org/@esbuild/linux-ia32/-/linux-ia32-0.27.3.tgz",
189
+
"integrity": "sha512-yGlQYjdxtLdh0a3jHjuwOrxQjOZYD/C9PfdbgJJF3TIZWnm/tMd/RcNiLngiu4iwcBAOezdnSLAwQDPqTmtTYg==",
190
+
"cpu": [
191
+
"ia32"
192
+
],
193
+
"dev": true,
194
+
"license": "MIT",
195
+
"optional": true,
196
+
"os": [
197
+
"linux"
198
+
],
199
+
"engines": {
200
+
"node": ">=18"
201
+
}
202
+
},
203
+
"node_modules/@esbuild/linux-loong64": {
204
+
"version": "0.27.3",
205
+
"resolved": "https://registry.npmjs.org/@esbuild/linux-loong64/-/linux-loong64-0.27.3.tgz",
206
+
"integrity": "sha512-WO60Sn8ly3gtzhyjATDgieJNet/KqsDlX5nRC5Y3oTFcS1l0KWba+SEa9Ja1GfDqSF1z6hif/SkpQJbL63cgOA==",
207
+
"cpu": [
208
+
"loong64"
209
+
],
210
+
"dev": true,
211
+
"license": "MIT",
212
+
"optional": true,
213
+
"os": [
214
+
"linux"
215
+
],
216
+
"engines": {
217
+
"node": ">=18"
218
+
}
219
+
},
220
+
"node_modules/@esbuild/linux-mips64el": {
221
+
"version": "0.27.3",
222
+
"resolved": "https://registry.npmjs.org/@esbuild/linux-mips64el/-/linux-mips64el-0.27.3.tgz",
223
+
"integrity": "sha512-APsymYA6sGcZ4pD6k+UxbDjOFSvPWyZhjaiPyl/f79xKxwTnrn5QUnXR5prvetuaSMsb4jgeHewIDCIWljrSxw==",
224
+
"cpu": [
225
+
"mips64el"
226
+
],
227
+
"dev": true,
228
+
"license": "MIT",
229
+
"optional": true,
230
+
"os": [
231
+
"linux"
232
+
],
233
+
"engines": {
234
+
"node": ">=18"
235
+
}
236
+
},
237
+
"node_modules/@esbuild/linux-ppc64": {
238
+
"version": "0.27.3",
239
+
"resolved": "https://registry.npmjs.org/@esbuild/linux-ppc64/-/linux-ppc64-0.27.3.tgz",
240
+
"integrity": "sha512-eizBnTeBefojtDb9nSh4vvVQ3V9Qf9Df01PfawPcRzJH4gFSgrObw+LveUyDoKU3kxi5+9RJTCWlj4FjYXVPEA==",
241
+
"cpu": [
242
+
"ppc64"
243
+
],
244
+
"dev": true,
245
+
"license": "MIT",
246
+
"optional": true,
247
+
"os": [
248
+
"linux"
249
+
],
250
+
"engines": {
251
+
"node": ">=18"
252
+
}
253
+
},
254
+
"node_modules/@esbuild/linux-riscv64": {
255
+
"version": "0.27.3",
256
+
"resolved": "https://registry.npmjs.org/@esbuild/linux-riscv64/-/linux-riscv64-0.27.3.tgz",
257
+
"integrity": "sha512-3Emwh0r5wmfm3ssTWRQSyVhbOHvqegUDRd0WhmXKX2mkHJe1SFCMJhagUleMq+Uci34wLSipf8Lagt4LlpRFWQ==",
258
+
"cpu": [
259
+
"riscv64"
260
+
],
261
+
"dev": true,
262
+
"license": "MIT",
263
+
"optional": true,
264
+
"os": [
265
+
"linux"
266
+
],
267
+
"engines": {
268
+
"node": ">=18"
269
+
}
270
+
},
271
+
"node_modules/@esbuild/linux-s390x": {
272
+
"version": "0.27.3",
273
+
"resolved": "https://registry.npmjs.org/@esbuild/linux-s390x/-/linux-s390x-0.27.3.tgz",
274
+
"integrity": "sha512-pBHUx9LzXWBc7MFIEEL0yD/ZVtNgLytvx60gES28GcWMqil8ElCYR4kvbV2BDqsHOvVDRrOxGySBM9Fcv744hw==",
275
+
"cpu": [
276
+
"s390x"
277
+
],
278
+
"dev": true,
279
+
"license": "MIT",
280
+
"optional": true,
281
+
"os": [
282
+
"linux"
283
+
],
284
+
"engines": {
285
+
"node": ">=18"
286
+
}
287
+
},
288
+
"node_modules/@esbuild/linux-x64": {
289
+
"version": "0.27.3",
290
+
"resolved": "https://registry.npmjs.org/@esbuild/linux-x64/-/linux-x64-0.27.3.tgz",
291
+
"integrity": "sha512-Czi8yzXUWIQYAtL/2y6vogER8pvcsOsk5cpwL4Gk5nJqH5UZiVByIY8Eorm5R13gq+DQKYg0+JyQoytLQas4dA==",
292
+
"cpu": [
293
+
"x64"
294
+
],
295
+
"dev": true,
296
+
"license": "MIT",
297
+
"optional": true,
298
+
"os": [
299
+
"linux"
300
+
],
301
+
"engines": {
302
+
"node": ">=18"
303
+
}
304
+
},
305
+
"node_modules/@esbuild/netbsd-arm64": {
306
+
"version": "0.27.3",
307
+
"resolved": "https://registry.npmjs.org/@esbuild/netbsd-arm64/-/netbsd-arm64-0.27.3.tgz",
308
+
"integrity": "sha512-sDpk0RgmTCR/5HguIZa9n9u+HVKf40fbEUt+iTzSnCaGvY9kFP0YKBWZtJaraonFnqef5SlJ8/TiPAxzyS+UoA==",
309
+
"cpu": [
310
+
"arm64"
311
+
],
312
+
"dev": true,
313
+
"license": "MIT",
314
+
"optional": true,
315
+
"os": [
316
+
"netbsd"
317
+
],
318
+
"engines": {
319
+
"node": ">=18"
320
+
}
321
+
},
322
+
"node_modules/@esbuild/netbsd-x64": {
323
+
"version": "0.27.3",
324
+
"resolved": "https://registry.npmjs.org/@esbuild/netbsd-x64/-/netbsd-x64-0.27.3.tgz",
325
+
"integrity": "sha512-P14lFKJl/DdaE00LItAukUdZO5iqNH7+PjoBm+fLQjtxfcfFE20Xf5CrLsmZdq5LFFZzb5JMZ9grUwvtVYzjiA==",
326
+
"cpu": [
327
+
"x64"
328
+
],
329
+
"dev": true,
330
+
"license": "MIT",
331
+
"optional": true,
332
+
"os": [
333
+
"netbsd"
334
+
],
335
+
"engines": {
336
+
"node": ">=18"
337
+
}
338
+
},
339
+
"node_modules/@esbuild/openbsd-arm64": {
340
+
"version": "0.27.3",
341
+
"resolved": "https://registry.npmjs.org/@esbuild/openbsd-arm64/-/openbsd-arm64-0.27.3.tgz",
342
+
"integrity": "sha512-AIcMP77AvirGbRl/UZFTq5hjXK+2wC7qFRGoHSDrZ5v5b8DK/GYpXW3CPRL53NkvDqb9D+alBiC/dV0Fb7eJcw==",
343
+
"cpu": [
344
+
"arm64"
345
+
],
346
+
"dev": true,
347
+
"license": "MIT",
348
+
"optional": true,
349
+
"os": [
350
+
"openbsd"
351
+
],
352
+
"engines": {
353
+
"node": ">=18"
354
+
}
355
+
},
356
+
"node_modules/@esbuild/openbsd-x64": {
357
+
"version": "0.27.3",
358
+
"resolved": "https://registry.npmjs.org/@esbuild/openbsd-x64/-/openbsd-x64-0.27.3.tgz",
359
+
"integrity": "sha512-DnW2sRrBzA+YnE70LKqnM3P+z8vehfJWHXECbwBmH/CU51z6FiqTQTHFenPlHmo3a8UgpLyH3PT+87OViOh1AQ==",
360
+
"cpu": [
361
+
"x64"
362
+
],
363
+
"dev": true,
364
+
"license": "MIT",
365
+
"optional": true,
366
+
"os": [
367
+
"openbsd"
368
+
],
369
+
"engines": {
370
+
"node": ">=18"
371
+
}
372
+
},
373
+
"node_modules/@esbuild/openharmony-arm64": {
374
+
"version": "0.27.3",
375
+
"resolved": "https://registry.npmjs.org/@esbuild/openharmony-arm64/-/openharmony-arm64-0.27.3.tgz",
376
+
"integrity": "sha512-NinAEgr/etERPTsZJ7aEZQvvg/A6IsZG/LgZy+81wON2huV7SrK3e63dU0XhyZP4RKGyTm7aOgmQk0bGp0fy2g==",
377
+
"cpu": [
378
+
"arm64"
379
+
],
380
+
"dev": true,
381
+
"license": "MIT",
382
+
"optional": true,
383
+
"os": [
384
+
"openharmony"
385
+
],
386
+
"engines": {
387
+
"node": ">=18"
388
+
}
389
+
},
390
+
"node_modules/@esbuild/sunos-x64": {
391
+
"version": "0.27.3",
392
+
"resolved": "https://registry.npmjs.org/@esbuild/sunos-x64/-/sunos-x64-0.27.3.tgz",
393
+
"integrity": "sha512-PanZ+nEz+eWoBJ8/f8HKxTTD172SKwdXebZ0ndd953gt1HRBbhMsaNqjTyYLGLPdoWHy4zLU7bDVJztF5f3BHA==",
394
+
"cpu": [
395
+
"x64"
396
+
],
397
+
"dev": true,
398
+
"license": "MIT",
399
+
"optional": true,
400
+
"os": [
401
+
"sunos"
402
+
],
403
+
"engines": {
404
+
"node": ">=18"
405
+
}
406
+
},
407
+
"node_modules/@esbuild/win32-arm64": {
408
+
"version": "0.27.3",
409
+
"resolved": "https://registry.npmjs.org/@esbuild/win32-arm64/-/win32-arm64-0.27.3.tgz",
410
+
"integrity": "sha512-B2t59lWWYrbRDw/tjiWOuzSsFh1Y/E95ofKz7rIVYSQkUYBjfSgf6oeYPNWHToFRr2zx52JKApIcAS/D5TUBnA==",
411
+
"cpu": [
412
+
"arm64"
413
+
],
414
+
"dev": true,
415
+
"license": "MIT",
416
+
"optional": true,
417
+
"os": [
418
+
"win32"
419
+
],
420
+
"engines": {
421
+
"node": ">=18"
422
+
}
423
+
},
424
+
"node_modules/@esbuild/win32-ia32": {
425
+
"version": "0.27.3",
426
+
"resolved": "https://registry.npmjs.org/@esbuild/win32-ia32/-/win32-ia32-0.27.3.tgz",
427
+
"integrity": "sha512-QLKSFeXNS8+tHW7tZpMtjlNb7HKau0QDpwm49u0vUp9y1WOF+PEzkU84y9GqYaAVW8aH8f3GcBck26jh54cX4Q==",
428
+
"cpu": [
429
+
"ia32"
430
+
],
431
+
"dev": true,
432
+
"license": "MIT",
433
+
"optional": true,
434
+
"os": [
435
+
"win32"
436
+
],
437
+
"engines": {
438
+
"node": ">=18"
439
+
}
440
+
},
441
+
"node_modules/@esbuild/win32-x64": {
442
+
"version": "0.27.3",
443
+
"resolved": "https://registry.npmjs.org/@esbuild/win32-x64/-/win32-x64-0.27.3.tgz",
444
+
"integrity": "sha512-4uJGhsxuptu3OcpVAzli+/gWusVGwZZHTlS63hh++ehExkVT8SgiEf7/uC/PclrPPkLhZqGgCTjd0VWLo6xMqA==",
445
+
"cpu": [
446
+
"x64"
447
+
],
448
+
"dev": true,
449
+
"license": "MIT",
450
+
"optional": true,
451
+
"os": [
452
+
"win32"
453
+
],
454
+
"engines": {
455
+
"node": ">=18"
456
+
}
457
+
},
458
+
"node_modules/@types/node": {
459
+
"version": "22.19.15",
460
+
"resolved": "https://registry.npmjs.org/@types/node/-/node-22.19.15.tgz",
461
+
"integrity": "sha512-F0R/h2+dsy5wJAUe3tAU6oqa2qbWY5TpNfL/RGmo1y38hiyO1w3x2jPtt76wmuaJI4DQnOBu21cNXQ2STIUUWg==",
462
+
"dev": true,
463
+
"license": "MIT",
464
+
"dependencies": {
465
+
"undici-types": "~6.21.0"
466
+
}
467
+
},
468
+
"node_modules/esbuild": {
469
+
"version": "0.27.3",
470
+
"resolved": "https://registry.npmjs.org/esbuild/-/esbuild-0.27.3.tgz",
471
+
"integrity": "sha512-8VwMnyGCONIs6cWue2IdpHxHnAjzxnw2Zr7MkVxB2vjmQ2ivqGFb4LEG3SMnv0Gb2F/G/2yA8zUaiL1gywDCCg==",
472
+
"dev": true,
473
+
"hasInstallScript": true,
474
+
"license": "MIT",
475
+
"bin": {
476
+
"esbuild": "bin/esbuild"
477
+
},
478
+
"engines": {
479
+
"node": ">=18"
480
+
},
481
+
"optionalDependencies": {
482
+
"@esbuild/aix-ppc64": "0.27.3",
483
+
"@esbuild/android-arm": "0.27.3",
484
+
"@esbuild/android-arm64": "0.27.3",
485
+
"@esbuild/android-x64": "0.27.3",
486
+
"@esbuild/darwin-arm64": "0.27.3",
487
+
"@esbuild/darwin-x64": "0.27.3",
488
+
"@esbuild/freebsd-arm64": "0.27.3",
489
+
"@esbuild/freebsd-x64": "0.27.3",
490
+
"@esbuild/linux-arm": "0.27.3",
491
+
"@esbuild/linux-arm64": "0.27.3",
492
+
"@esbuild/linux-ia32": "0.27.3",
493
+
"@esbuild/linux-loong64": "0.27.3",
494
+
"@esbuild/linux-mips64el": "0.27.3",
495
+
"@esbuild/linux-ppc64": "0.27.3",
496
+
"@esbuild/linux-riscv64": "0.27.3",
497
+
"@esbuild/linux-s390x": "0.27.3",
498
+
"@esbuild/linux-x64": "0.27.3",
499
+
"@esbuild/netbsd-arm64": "0.27.3",
500
+
"@esbuild/netbsd-x64": "0.27.3",
501
+
"@esbuild/openbsd-arm64": "0.27.3",
502
+
"@esbuild/openbsd-x64": "0.27.3",
503
+
"@esbuild/openharmony-arm64": "0.27.3",
504
+
"@esbuild/sunos-x64": "0.27.3",
505
+
"@esbuild/win32-arm64": "0.27.3",
506
+
"@esbuild/win32-ia32": "0.27.3",
507
+
"@esbuild/win32-x64": "0.27.3"
508
+
}
509
+
},
510
+
"node_modules/fsevents": {
511
+
"version": "2.3.3",
512
+
"resolved": "https://registry.npmjs.org/fsevents/-/fsevents-2.3.3.tgz",
513
+
"integrity": "sha512-5xoDfX+fL7faATnagmWPpbFtwh/R77WmMMqqHGS65C3vvB0YHrgF+B1YmZ3441tMj5n63k0212XNoJwzlhffQw==",
514
+
"dev": true,
515
+
"hasInstallScript": true,
516
+
"license": "MIT",
517
+
"optional": true,
518
+
"os": [
519
+
"darwin"
520
+
],
521
+
"engines": {
522
+
"node": "^8.16.0 || ^10.6.0 || >=11.0.0"
523
+
}
524
+
},
525
+
"node_modules/get-tsconfig": {
526
+
"version": "4.13.6",
527
+
"resolved": "https://registry.npmjs.org/get-tsconfig/-/get-tsconfig-4.13.6.tgz",
528
+
"integrity": "sha512-shZT/QMiSHc/YBLxxOkMtgSid5HFoauqCE3/exfsEcwg1WkeqjG+V40yBbBrsD+jW2HDXcs28xOfcbm2jI8Ddw==",
529
+
"dev": true,
530
+
"license": "MIT",
531
+
"dependencies": {
532
+
"resolve-pkg-maps": "^1.0.0"
533
+
},
534
+
"funding": {
535
+
"url": "https://github.com/privatenumber/get-tsconfig?sponsor=1"
536
+
}
537
+
},
538
+
"node_modules/resolve-pkg-maps": {
539
+
"version": "1.0.0",
540
+
"resolved": "https://registry.npmjs.org/resolve-pkg-maps/-/resolve-pkg-maps-1.0.0.tgz",
541
+
"integrity": "sha512-seS2Tj26TBVOC2NIc2rOe2y2ZO7efxITtLZcGSOnHHNOQ7CkiUBfw0Iw2ck6xkIhPwLhKNLS8BO+hEpngQlqzw==",
542
+
"dev": true,
543
+
"license": "MIT",
544
+
"funding": {
545
+
"url": "https://github.com/privatenumber/resolve-pkg-maps?sponsor=1"
546
+
}
547
+
},
548
+
"node_modules/tsx": {
549
+
"version": "4.21.0",
550
+
"resolved": "https://registry.npmjs.org/tsx/-/tsx-4.21.0.tgz",
551
+
"integrity": "sha512-5C1sg4USs1lfG0GFb2RLXsdpXqBSEhAaA/0kPL01wxzpMqLILNxIxIOKiILz+cdg/pLnOUxFYOR5yhHU666wbw==",
552
+
"dev": true,
553
+
"license": "MIT",
554
+
"dependencies": {
555
+
"esbuild": "~0.27.0",
556
+
"get-tsconfig": "^4.7.5"
557
+
},
558
+
"bin": {
559
+
"tsx": "dist/cli.mjs"
560
+
},
561
+
"engines": {
562
+
"node": ">=18.0.0"
563
+
},
564
+
"optionalDependencies": {
565
+
"fsevents": "~2.3.3"
566
+
}
567
+
},
568
+
"node_modules/typescript": {
569
+
"version": "5.9.3",
570
+
"resolved": "https://registry.npmjs.org/typescript/-/typescript-5.9.3.tgz",
571
+
"integrity": "sha512-jl1vZzPDinLr9eUt3J/t7V6FgNEw9QjvBPdysz9KfQDD41fQrC2Y4vKQdiaUpFT4bXlb1RHhLpp8wtm6M5TgSw==",
572
+
"dev": true,
573
+
"license": "Apache-2.0",
574
+
"bin": {
575
+
"tsc": "bin/tsc",
576
+
"tsserver": "bin/tsserver"
577
+
},
578
+
"engines": {
579
+
"node": ">=14.17"
580
+
}
581
+
},
582
+
"node_modules/undici-types": {
583
+
"version": "6.21.0",
584
+
"resolved": "https://registry.npmjs.org/undici-types/-/undici-types-6.21.0.tgz",
585
+
"integrity": "sha512-iwDZqg0QAGrg9Rav5H4n0M64c3mkR59cJ6wQp+7C4nI0gsmExaedaYLNO44eT4AtBBwjbTiGPMlt2Md0T9H9JQ==",
586
+
"dev": true,
587
+
"license": "MIT"
588
+
}
589
+
}
590
+
}
+18
package.json
+18
package.json
···
1
+
{
2
+
"name": "localcode",
3
+
"version": "2.0.0",
4
+
"type": "module",
5
+
"bin": {
6
+
"localcode": "dist/main.js"
7
+
},
8
+
"scripts": {
9
+
"dev": "tsx src/main.ts",
10
+
"build": "tsc",
11
+
"start": "node dist/main.js"
12
+
},
13
+
"devDependencies": {
14
+
"@types/node": "^22",
15
+
"tsx": "^4",
16
+
"typescript": "^5"
17
+
}
18
+
}
+406
setup.sh
+406
setup.sh
···
1
+
#!/usr/bin/env bash
2
+
set -euo pipefail
3
+
4
+
# =============================================================================
5
+
# Local AI Coding Environment Setup for macOS Apple Silicon
6
+
# llama.cpp + Qwen 2.5 Coder 32B (chat) + Qwen 2.5 Coder 1.5B (autocomplete)
7
+
# + Aider (terminal coding agent) or OpenCode
8
+
# =============================================================================
9
+
10
+
BOLD="\033[1m"
11
+
GREEN="\033[0;32m"
12
+
YELLOW="\033[1;33m"
13
+
RED="\033[0;31m"
14
+
RESET="\033[0m"
15
+
16
+
MODELS_DIR="$HOME/.local/share/llama-models"
17
+
CHAT_MODEL_URL="https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct-GGUF/resolve/main/qwen2.5-coder-32b-instruct-q4_k_m.gguf"
18
+
CHAT_MODEL_FILE="qwen2.5-coder-32b-instruct-q4_k_m.gguf"
19
+
AUTOCOMPLETE_MODEL_URL="https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF/resolve/main/qwen2.5-coder-1.5b-instruct-q4_k_m.gguf"
20
+
AUTOCOMPLETE_MODEL_FILE="qwen2.5-coder-1.5b-instruct-q4_k_m.gguf"
21
+
22
+
CHAT_PORT=8080
23
+
AUTOCOMPLETE_PORT=8081
24
+
25
+
AIDER_CONFIG_DIR="$HOME/.aider"
26
+
AIDER_CONFIG_FILE="$AIDER_CONFIG_DIR/aider.conf.yml"
27
+
28
+
log() { echo -e "${GREEN}${BOLD}[✓]${RESET} $1"; }
29
+
warn() { echo -e "${YELLOW}${BOLD}[!]${RESET} $1"; }
30
+
err() { echo -e "${RED}${BOLD}[✗]${RESET} $1"; exit 1; }
31
+
32
+
# -----------------------------------------------------------------------------
33
+
# Pre-flight checks
34
+
# -----------------------------------------------------------------------------
35
+
echo -e "\n${BOLD}🔧 Local AI Coding Environment Installer (llama.cpp)${RESET}\n"
36
+
37
+
[[ "$(uname)" == "Darwin" ]] || err "This script is for macOS only."
38
+
[[ "$(uname -m)" == "arm64" ]] || warn "Not running on Apple Silicon — performance may vary."
39
+
40
+
MEM_GB=$(( $(sysctl -n hw.memsize) / 1073741824 ))
41
+
if (( MEM_GB < 32 )); then
42
+
warn "You have ${MEM_GB}GB RAM. The 32B model needs ~20GB; you may experience swapping."
43
+
fi
44
+
45
+
# -----------------------------------------------------------------------------
46
+
# 1. Install Homebrew (if missing)
47
+
# -----------------------------------------------------------------------------
48
+
if ! command -v brew &>/dev/null; then
49
+
log "Installing Homebrew..."
50
+
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
51
+
eval "$(/opt/homebrew/bin/brew shellenv)"
52
+
else
53
+
log "Homebrew already installed."
54
+
fi
55
+
56
+
# -----------------------------------------------------------------------------
57
+
# 2. Build / Install llama.cpp
58
+
# -----------------------------------------------------------------------------
59
+
if ! command -v llama-server &>/dev/null; then
60
+
log "Installing llama.cpp via Homebrew..."
61
+
brew install llama.cpp
62
+
else
63
+
log "llama.cpp already installed."
64
+
fi
65
+
66
+
# Verify Metal support
67
+
if llama-server --help 2>&1 | grep -qi metal; then
68
+
log "Metal (GPU) acceleration available."
69
+
else
70
+
warn "Metal flag not detected — model will run on CPU only."
71
+
fi
72
+
73
+
# -----------------------------------------------------------------------------
74
+
# 3. Download Qwen GGUF models from HuggingFace
75
+
# -----------------------------------------------------------------------------
76
+
mkdir -p "$MODELS_DIR"
77
+
78
+
download_model() {
79
+
local url="$1" file="$2"
80
+
if [ -f "$MODELS_DIR/$file" ]; then
81
+
log "Model already downloaded: $file"
82
+
else
83
+
log "Downloading $file (this may take a while)..."
84
+
curl -L --progress-bar -o "$MODELS_DIR/$file" "$url"
85
+
log "Downloaded: $file"
86
+
fi
87
+
}
88
+
89
+
download_model "$CHAT_MODEL_URL" "$CHAT_MODEL_FILE"
90
+
download_model "$AUTOCOMPLETE_MODEL_URL" "$AUTOCOMPLETE_MODEL_FILE"
91
+
92
+
# -----------------------------------------------------------------------------
93
+
# 4. Install Python & Aider
94
+
# -----------------------------------------------------------------------------
95
+
if ! command -v jq &>/dev/null; then
96
+
log "Installing jq..."
97
+
brew install jq
98
+
else
99
+
log "jq already installed."
100
+
fi
101
+
102
+
if ! command -v python3 &>/dev/null; then
103
+
log "Installing Python 3..."
104
+
brew install python@3.12
105
+
fi
106
+
107
+
if ! command -v pipx &>/dev/null; then
108
+
log "Installing pipx..."
109
+
brew install pipx
110
+
pipx ensurepath
111
+
fi
112
+
113
+
if ! command -v aider &>/dev/null; then
114
+
log "Installing Aider..."
115
+
pipx install aider-chat
116
+
else
117
+
log "Aider already installed. Upgrading..."
118
+
pipx upgrade aider-chat
119
+
fi
120
+
121
+
# -----------------------------------------------------------------------------
122
+
# 5. Create llama.cpp server launcher scripts
123
+
# -----------------------------------------------------------------------------
124
+
LAUNCH_DIR="$HOME/.local/bin"
125
+
mkdir -p "$LAUNCH_DIR"
126
+
127
+
# --- Chat server launcher ---
128
+
cat > "$LAUNCH_DIR/llama-chat-server" << SCRIPT
129
+
#!/usr/bin/env bash
130
+
# Start llama.cpp server with Qwen 2.5 Coder 32B for chat
131
+
# Exposed as OpenAI-compatible API on port ${CHAT_PORT}
132
+
133
+
MODEL="$MODELS_DIR/$CHAT_MODEL_FILE"
134
+
135
+
exec llama-server \\
136
+
--model "\$MODEL" \\
137
+
--port ${CHAT_PORT} \\
138
+
--host 127.0.0.1 \\
139
+
--ctx-size 16384 \\
140
+
--n-gpu-layers 99 \\
141
+
--threads \$(sysctl -n hw.perflevel0.logicalcpu 2>/dev/null || echo 4) \\
142
+
--mlock \\
143
+
"\$@"
144
+
SCRIPT
145
+
chmod +x "$LAUNCH_DIR/llama-chat-server"
146
+
147
+
# --- Autocomplete server launcher ---
148
+
cat > "$LAUNCH_DIR/llama-complete-server" << SCRIPT
149
+
#!/usr/bin/env bash
150
+
# Start llama.cpp server with Qwen 2.5 Coder 1.5B for autocomplete
151
+
# Exposed as OpenAI-compatible API on port ${AUTOCOMPLETE_PORT}
152
+
153
+
MODEL="$MODELS_DIR/$AUTOCOMPLETE_MODEL_FILE"
154
+
155
+
exec llama-server \\
156
+
--model "\$MODEL" \\
157
+
--port ${AUTOCOMPLETE_PORT} \\
158
+
--host 127.0.0.1 \\
159
+
--ctx-size 4096 \\
160
+
--n-gpu-layers 99 \\
161
+
--threads \$(sysctl -n hw.perflevel0.logicalcpu 2>/dev/null || echo 4) \\
162
+
--mlock \\
163
+
"\$@"
164
+
SCRIPT
165
+
chmod +x "$LAUNCH_DIR/llama-complete-server"
166
+
167
+
# --- Combined server manager ---
168
+
cat > "$LAUNCH_DIR/llama-start" << 'SCRIPT'
169
+
#!/usr/bin/env bash
170
+
# Start both llama.cpp servers (chat + autocomplete)
171
+
172
+
BOLD="\033[1m"; GREEN="\033[0;32m"; RED="\033[0;31m"; RESET="\033[0m"
173
+
CHAT_PID="" COMPLETE_PID=""
174
+
175
+
cleanup() {
176
+
echo -e "\n${RED}Shutting down servers...${RESET}"
177
+
[ -n "$CHAT_PID" ] && kill "$CHAT_PID" 2>/dev/null
178
+
[ -n "$COMPLETE_PID" ] && kill "$COMPLETE_PID" 2>/dev/null
179
+
wait 2>/dev/null
180
+
echo -e "${GREEN}Done.${RESET}"
181
+
exit 0
182
+
}
183
+
trap cleanup SIGINT SIGTERM
184
+
185
+
echo -e "${BOLD}Starting llama.cpp servers...${RESET}\n"
186
+
187
+
echo -e "${GREEN}[1/2]${RESET} Chat model (32B) on :8080..."
188
+
llama-chat-server &>/tmp/llama-chat.log &
189
+
CHAT_PID=$!
190
+
191
+
echo -e "${GREEN}[2/2]${RESET} Autocomplete model (1.5B) on :8081..."
192
+
llama-complete-server &>/tmp/llama-complete.log &
193
+
COMPLETE_PID=$!
194
+
195
+
# Wait for servers to be ready
196
+
echo -ne "\nWaiting for servers..."
197
+
for i in $(seq 1 60); do
198
+
CHAT_OK=$(curl -s -o /dev/null -w "%{http_code}" http://127.0.0.1:8080/health 2>/dev/null || true)
199
+
COMP_OK=$(curl -s -o /dev/null -w "%{http_code}" http://127.0.0.1:8081/health 2>/dev/null || true)
200
+
if [[ "$CHAT_OK" == "200" && "$COMP_OK" == "200" ]]; then
201
+
echo -e " ${GREEN}ready!${RESET}"
202
+
break
203
+
fi
204
+
echo -n "."
205
+
sleep 2
206
+
done
207
+
208
+
echo ""
209
+
echo -e "${BOLD}Servers running:${RESET}"
210
+
echo -e " Chat (32B): http://127.0.0.1:8080"
211
+
echo -e " Autocomplete (1.5B): http://127.0.0.1:8081"
212
+
echo -e " Logs: /tmp/llama-chat.log, /tmp/llama-complete.log"
213
+
echo -e "\n Press Ctrl+C to stop both servers.\n"
214
+
215
+
wait
216
+
SCRIPT
217
+
chmod +x "$LAUNCH_DIR/llama-start"
218
+
219
+
# --- Stop servers ---
220
+
cat > "$LAUNCH_DIR/llama-stop" << 'SCRIPT'
221
+
#!/usr/bin/env bash
222
+
# Stop all running llama-server processes
223
+
pkill -f "llama-server" 2>/dev/null && echo "Servers stopped." || echo "No servers running."
224
+
SCRIPT
225
+
chmod +x "$LAUNCH_DIR/llama-stop"
226
+
227
+
# -----------------------------------------------------------------------------
228
+
# 6. Configure Aider to use llama.cpp OpenAI-compatible API
229
+
# -----------------------------------------------------------------------------
230
+
mkdir -p "$AIDER_CONFIG_DIR"
231
+
232
+
cat > "$AIDER_CONFIG_FILE" << 'EOF'
233
+
# =============================================================================
234
+
# Aider Configuration — Qwen 2.5 Coder via llama.cpp
235
+
# =============================================================================
236
+
237
+
# Point Aider at llama.cpp's OpenAI-compatible endpoint
238
+
# The model name can be anything — llama.cpp ignores it and uses the loaded model
239
+
model: openai/qwen2.5-coder-32b
240
+
241
+
# Architect mode for better code planning
242
+
architect: true
243
+
editor-model: openai/qwen2.5-coder-32b
244
+
245
+
# Git integration
246
+
auto-commits: true
247
+
dirty-commits: true
248
+
attribute-author: false
249
+
attribute-committer: false
250
+
251
+
# UI preferences
252
+
pretty: true
253
+
stream: true
254
+
dark-mode: true
255
+
256
+
# Code style
257
+
code-theme: monokai
258
+
show-diffs: true
259
+
260
+
# Disable analytics
261
+
analytics-disable: true
262
+
EOF
263
+
264
+
# Environment file for API base URL
265
+
cat > "$AIDER_CONFIG_DIR/.env" << 'EOF'
266
+
# llama.cpp serves an OpenAI-compatible API — no real key needed
267
+
OPENAI_API_KEY=sk-not-needed
268
+
OPENAI_API_BASE=http://127.0.0.1:8080/v1
269
+
EOF
270
+
271
+
log "Aider config written to ${AIDER_CONFIG_FILE}"
272
+
log "Aider env written to ${AIDER_CONFIG_DIR}/.env"
273
+
274
+
# -----------------------------------------------------------------------------
275
+
# 7. Create main launcher: ai-code
276
+
# -----------------------------------------------------------------------------
277
+
cat > "$LAUNCH_DIR/ai-code" << 'SCRIPT'
278
+
#!/usr/bin/env bash
279
+
# Launch Aider with local Qwen 2.5 Coder 32B via llama.cpp
280
+
# Usage: ai-code [directory] [aider flags...]
281
+
#
282
+
# Starts llama.cpp servers automatically if not already running.
283
+
284
+
BOLD="\033[1m"; GREEN="\033[0;32m"; RESET="\033[0m"
285
+
286
+
# Check if chat server is running
287
+
if ! curl -s http://127.0.0.1:8080/health &>/dev/null; then
288
+
echo -e "${BOLD}Starting llama.cpp chat server...${RESET}"
289
+
llama-chat-server &>/tmp/llama-chat.log &
290
+
echo -n "Waiting for model to load"
291
+
for i in $(seq 1 120); do
292
+
if curl -s http://127.0.0.1:8080/health &>/dev/null; then
293
+
echo -e " ${GREEN}ready!${RESET}"
294
+
break
295
+
fi
296
+
echo -n "."
297
+
sleep 2
298
+
done
299
+
fi
300
+
301
+
DIR="${1:-.}"
302
+
shift 2>/dev/null || true
303
+
cd "$DIR" || exit 1
304
+
305
+
# Initialize git repo if needed
306
+
if [ ! -d .git ]; then
307
+
echo "Initializing git repo..."
308
+
git init
309
+
git add -A
310
+
git commit -m "Initial commit (before AI edits)" --allow-empty
311
+
fi
312
+
313
+
# Source the env file for API config
314
+
export $(grep -v '^#' "$HOME/.aider/.env" | xargs)
315
+
316
+
exec aider "$@"
317
+
SCRIPT
318
+
chmod +x "$LAUNCH_DIR/ai-code"
319
+
320
+
# Quick question mode
321
+
cat > "$LAUNCH_DIR/ai-ask" << 'SCRIPT'
322
+
#!/usr/bin/env bash
323
+
# Quick coding Q&A — no file editing
324
+
if ! curl -s http://127.0.0.1:8080/health &>/dev/null; then
325
+
llama-chat-server &>/tmp/llama-chat.log &
326
+
sleep 5
327
+
fi
328
+
export $(grep -v '^#' "$HOME/.aider/.env" | xargs)
329
+
if [ -n "$1" ]; then
330
+
exec aider --no-auto-commits --message "$*"
331
+
else
332
+
exec aider --no-auto-commits
333
+
fi
334
+
SCRIPT
335
+
chmod +x "$LAUNCH_DIR/ai-ask"
336
+
337
+
# Pipe mode using llama.cpp CLI directly
338
+
cat > "$LAUNCH_DIR/ai-pipe" << SCRIPT
339
+
#!/usr/bin/env bash
340
+
# Pipe code through llama.cpp
341
+
# Usage: cat main.py | ai-pipe "add error handling"
342
+
343
+
PROMPT="\${1:-Improve this code}"
344
+
INPUT=\$(cat)
345
+
346
+
curl -s http://127.0.0.1:8080/v1/chat/completions \\
347
+
-H "Content-Type: application/json" \\
348
+
-d "\$(jq -n --arg p "\$PROMPT" --arg c "\$INPUT" '{
349
+
model: "qwen",
350
+
messages: [
351
+
{role: "system", content: "You are an expert programmer. Output only code, no explanations."},
352
+
{role: "user", content: ("\$p\n\n```\n" + \$c + "\n```")}
353
+
],
354
+
stream: false
355
+
}')" | jq -r '.choices[0].message.content'
356
+
SCRIPT
357
+
chmod +x "$LAUNCH_DIR/ai-pipe"
358
+
359
+
# -----------------------------------------------------------------------------
360
+
# 8. Shell integration
361
+
# -----------------------------------------------------------------------------
362
+
SHELL_RC=""
363
+
case "$SHELL" in
364
+
*/zsh) SHELL_RC="$HOME/.zshrc" ;;
365
+
*/bash) SHELL_RC="$HOME/.bashrc" ;;
366
+
*) SHELL_RC="$HOME/.profile" ;;
367
+
esac
368
+
369
+
if ! grep -q '.local/bin' "$SHELL_RC" 2>/dev/null; then
370
+
echo '' >> "$SHELL_RC"
371
+
echo '# Local AI coding tools' >> "$SHELL_RC"
372
+
echo 'export PATH="$HOME/.local/bin:$PATH"' >> "$SHELL_RC"
373
+
log "Added ~/.local/bin to PATH in ${SHELL_RC}"
374
+
fi
375
+
376
+
# -----------------------------------------------------------------------------
377
+
# Done!
378
+
# -----------------------------------------------------------------------------
379
+
echo ""
380
+
echo -e "${GREEN}${BOLD}═══════════════════════════════════════════════════${RESET}"
381
+
echo -e "${GREEN}${BOLD} ✅ Setup complete!${RESET}"
382
+
echo -e "${GREEN}${BOLD}═══════════════════════════════════════════════════${RESET}"
383
+
echo ""
384
+
echo -e " ${BOLD}Models downloaded to:${RESET} ${MODELS_DIR}"
385
+
echo -e " Chat: ${CHAT_MODEL_FILE} (~20GB)"
386
+
echo -e " Autocomplete: ${AUTOCOMPLETE_MODEL_FILE} (~1.2GB)"
387
+
echo ""
388
+
echo -e " ${BOLD}Commands available${RESET} (restart your shell first):"
389
+
echo ""
390
+
echo -e " ${BOLD}llama-start${RESET} Start both llama.cpp servers"
391
+
echo -e " ${BOLD}llama-stop${RESET} Stop all llama.cpp servers"
392
+
echo ""
393
+
echo -e " ${BOLD}ai-code${RESET} [dir] Full coding agent (auto-starts server)"
394
+
echo -e " cd into a project and run 'ai-code .'"
395
+
echo ""
396
+
echo -e " ${BOLD}ai-ask${RESET} \"question\" Quick coding Q&A, no file edits"
397
+
echo ""
398
+
echo -e " ${BOLD}ai-pipe${RESET} \"prompt\" Pipe code through the model"
399
+
echo -e " cat file.py | ai-pipe \"add types\""
400
+
echo ""
401
+
echo -e " ${BOLD}Config:${RESET} ${AIDER_CONFIG_FILE}"
402
+
echo -e " ${BOLD}API env:${RESET} ${AIDER_CONFIG_DIR}/.env"
403
+
echo -e " ${BOLD}Server logs:${RESET} /tmp/llama-chat.log, /tmp/llama-complete.log"
404
+
echo ""
405
+
echo -e " Run ${BOLD}source ${SHELL_RC}${RESET} or open a new terminal to get started."
406
+
echo ""
+275
src/commands/bench.ts
+275
src/commands/bench.ts
···
1
+
import { readFileSync, writeFileSync, mkdirSync } from "node:fs";
2
+
import { join, dirname } from "node:path";
3
+
import { homedir } from "node:os";
4
+
import { performance } from "node:perf_hooks";
5
+
import { CHAT_PORT } from "../config.js";
6
+
import { getActiveChatModel } from "../runtime-config.js";
7
+
import { log, warn, err } from "../log.js";
8
+
import type { ModelDef } from "../registry/models.js";
9
+
10
+
const BOLD = "\x1b[1m";
11
+
const DIM = "\x1b[2m";
12
+
const RESET = "\x1b[0m";
13
+
14
+
const BENCHMARKS_PATH = join(
15
+
homedir(),
16
+
".config",
17
+
"localcode",
18
+
"benchmarks.json",
19
+
);
20
+
21
+
interface BenchPrompt {
22
+
label: string;
23
+
system: string;
24
+
user: string;
25
+
}
26
+
27
+
const PROMPTS: BenchPrompt[] = [
28
+
{
29
+
label: "fizzbuzz",
30
+
system: "You are an expert programmer.",
31
+
user: "Write a fizzbuzz function in Python.",
32
+
},
33
+
{
34
+
label: "BST class",
35
+
system: "You are an expert programmer.",
36
+
user: "Write a binary search tree implementation in TypeScript with insert, delete, and search methods.",
37
+
},
38
+
{
39
+
label: "code review",
40
+
system: "You are an expert code reviewer.",
41
+
user: `Review this code and suggest improvements:
42
+
43
+
function processData(data) {
44
+
var result = [];
45
+
for (var i = 0; i < data.length; i++) {
46
+
if (data[i].active == true) {
47
+
var item = {};
48
+
item.name = data[i].firstName + " " + data[i].lastName;
49
+
item.email = data[i].email;
50
+
item.score = data[i].points / data[i].maxPoints * 100;
51
+
if (item.score >= 90) {
52
+
item.grade = "A";
53
+
} else if (item.score >= 80) {
54
+
item.grade = "B";
55
+
} else if (item.score >= 70) {
56
+
item.grade = "C";
57
+
} else if (item.score >= 60) {
58
+
item.grade = "D";
59
+
} else {
60
+
item.grade = "F";
61
+
}
62
+
result.push(item);
63
+
}
64
+
}
65
+
result.sort(function(a, b) { return b.score - a.score; });
66
+
return result;
67
+
}`,
68
+
},
69
+
];
70
+
71
+
interface PromptResult {
72
+
label: string;
73
+
promptTokens: number;
74
+
completionTokens: number;
75
+
elapsedMs: number;
76
+
tokensPerSec: number;
77
+
}
78
+
79
+
interface BenchmarkEntry {
80
+
timestamp: string;
81
+
model: string;
82
+
modelName: string;
83
+
results: PromptResult[];
84
+
avgTokPerSec: number;
85
+
}
86
+
87
+
async function checkHealth(port: number): Promise<boolean> {
88
+
try {
89
+
const res = await fetch(`http://127.0.0.1:${port}/health`);
90
+
return res.ok;
91
+
} catch {
92
+
return false;
93
+
}
94
+
}
95
+
96
+
async function runPrompt(
97
+
port: number,
98
+
prompt: BenchPrompt,
99
+
): Promise<PromptResult> {
100
+
const body = JSON.stringify({
101
+
model: "qwen",
102
+
messages: [
103
+
{ role: "system", content: prompt.system },
104
+
{ role: "user", content: prompt.user },
105
+
],
106
+
stream: false,
107
+
});
108
+
109
+
const start = performance.now();
110
+
const res = await fetch(
111
+
`http://127.0.0.1:${port}/v1/chat/completions`,
112
+
{
113
+
method: "POST",
114
+
headers: { "Content-Type": "application/json" },
115
+
body,
116
+
},
117
+
);
118
+
119
+
if (!res.ok) {
120
+
const text = await res.text();
121
+
throw new Error(`Server returned ${res.status}: ${text}`);
122
+
}
123
+
124
+
const elapsed = performance.now() - start;
125
+
const data = (await res.json()) as {
126
+
usage?: { prompt_tokens?: number; completion_tokens?: number };
127
+
};
128
+
129
+
const promptTokens = data.usage?.prompt_tokens ?? 0;
130
+
const completionTokens = data.usage?.completion_tokens ?? 0;
131
+
const tokPerSec =
132
+
completionTokens > 0 ? completionTokens / (elapsed / 1000) : 0;
133
+
134
+
return {
135
+
label: prompt.label,
136
+
promptTokens,
137
+
completionTokens,
138
+
elapsedMs: elapsed,
139
+
tokensPerSec: tokPerSec,
140
+
};
141
+
}
142
+
143
+
function saveBenchmark(entry: BenchmarkEntry): void {
144
+
let history: BenchmarkEntry[] = [];
145
+
try {
146
+
history = JSON.parse(
147
+
readFileSync(BENCHMARKS_PATH, "utf-8"),
148
+
) as BenchmarkEntry[];
149
+
} catch {
150
+
// No existing file
151
+
}
152
+
history.push(entry);
153
+
mkdirSync(dirname(BENCHMARKS_PATH), { recursive: true });
154
+
writeFileSync(BENCHMARKS_PATH, JSON.stringify(history, null, 2) + "\n");
155
+
}
156
+
157
+
function printResults(model: ModelDef, results: PromptResult[]): void {
158
+
console.log("");
159
+
console.log(`${BOLD}Model:${RESET} ${model.name} (${model.file})`);
160
+
console.log(`${BOLD}Port:${RESET} ${CHAT_PORT}`);
161
+
console.log("");
162
+
163
+
// Table header
164
+
const hdr = [
165
+
"Prompt".padEnd(16),
166
+
"Prompt Tok".padStart(10),
167
+
"Compl Tok".padStart(10),
168
+
"Time (s)".padStart(10),
169
+
"Tok/s".padStart(8),
170
+
].join(" ");
171
+
console.log(` ${BOLD}${hdr}${RESET}`);
172
+
console.log(` ${"─".repeat(hdr.length)}`);
173
+
174
+
for (const r of results) {
175
+
const row = [
176
+
r.label.padEnd(16),
177
+
String(r.promptTokens).padStart(10),
178
+
String(r.completionTokens).padStart(10),
179
+
(r.elapsedMs / 1000).toFixed(1).padStart(10),
180
+
r.tokensPerSec.toFixed(1).padStart(8),
181
+
].join(" ");
182
+
console.log(` ${row}`);
183
+
}
184
+
185
+
const avgTokSec =
186
+
results.reduce((s, r) => s + r.tokensPerSec, 0) / results.length;
187
+
console.log("");
188
+
console.log(` ${BOLD}Average: ${avgTokSec.toFixed(1)} tok/s${RESET}`);
189
+
console.log("");
190
+
}
191
+
192
+
function printHistory(): void {
193
+
let history: BenchmarkEntry[] = [];
194
+
try {
195
+
history = JSON.parse(
196
+
readFileSync(BENCHMARKS_PATH, "utf-8"),
197
+
) as BenchmarkEntry[];
198
+
} catch {
199
+
console.log("No benchmark history found.");
200
+
return;
201
+
}
202
+
203
+
if (history.length === 0) {
204
+
console.log("No benchmark history found.");
205
+
return;
206
+
}
207
+
208
+
console.log(`\n${BOLD}Benchmark History:${RESET}\n`);
209
+
const hdr = [
210
+
"Date".padEnd(20),
211
+
"Model".padEnd(24),
212
+
"Avg Tok/s".padStart(10),
213
+
].join(" ");
214
+
console.log(` ${BOLD}${hdr}${RESET}`);
215
+
console.log(` ${"─".repeat(hdr.length)}`);
216
+
217
+
for (const entry of history) {
218
+
const date = entry.timestamp.replace("T", " ").slice(0, 19);
219
+
const row = [
220
+
date.padEnd(20),
221
+
entry.modelName.padEnd(24),
222
+
entry.avgTokPerSec.toFixed(1).padStart(10),
223
+
].join(" ");
224
+
console.log(` ${row}`);
225
+
}
226
+
console.log("");
227
+
}
228
+
229
+
export async function runBench(args: string[]): Promise<void> {
230
+
if (args.includes("--history")) {
231
+
printHistory();
232
+
return;
233
+
}
234
+
235
+
const healthy = await checkHealth(CHAT_PORT);
236
+
if (!healthy) {
237
+
err(
238
+
`Chat server not running on port ${CHAT_PORT}.\nStart it with: llama-start`,
239
+
);
240
+
}
241
+
242
+
const model = getActiveChatModel();
243
+
log(`Benchmarking ${model.name} on port ${CHAT_PORT}...`);
244
+
console.log(`${DIM}Running ${PROMPTS.length} prompts (this may take a minute)...${RESET}`);
245
+
246
+
const results: PromptResult[] = [];
247
+
for (const prompt of PROMPTS) {
248
+
process.stdout.write(` ${prompt.label}...`);
249
+
try {
250
+
const result = await runPrompt(CHAT_PORT, prompt);
251
+
results.push(result);
252
+
console.log(` ${result.tokensPerSec.toFixed(1)} tok/s`);
253
+
} catch (e) {
254
+
console.log(` FAILED: ${e instanceof Error ? e.message : e}`);
255
+
}
256
+
}
257
+
258
+
if (results.length === 0) {
259
+
err("All prompts failed.");
260
+
}
261
+
262
+
printResults(model, results);
263
+
264
+
// Save to history
265
+
const avgTokPerSec =
266
+
results.reduce((s, r) => s + r.tokensPerSec, 0) / results.length;
267
+
saveBenchmark({
268
+
timestamp: new Date().toISOString(),
269
+
model: model.id,
270
+
modelName: model.name,
271
+
results,
272
+
avgTokPerSec,
273
+
});
274
+
log(`Results saved to ${BENCHMARKS_PATH}`);
275
+
}
+102
src/commands/models.ts
+102
src/commands/models.ts
···
1
+
import { existsSync } from "node:fs";
2
+
import { join } from "node:path";
3
+
import { MODELS, getChatModels, getAutocompleteModels, getModelById } from "../registry/models.js";
4
+
import { loadConfig, saveConfig, getActiveChatModel, getActiveAutocompleteModel } from "../runtime-config.js";
5
+
import { createLauncherScripts } from "../steps/scripts.js";
6
+
import { writeTuiConfig } from "../steps/aider-config.js";
7
+
import { MODELS_DIR } from "../config.js";
8
+
import { log, err } from "../log.js";
9
+
import { runPassthrough } from "../util.js";
10
+
import { mkdirSync } from "node:fs";
11
+
12
+
const BOLD = "\x1b[1m";
13
+
const GREEN = "\x1b[0;32m";
14
+
const DIM = "\x1b[2m";
15
+
const RESET = "\x1b[0m";
16
+
17
+
function isDownloaded(file: string): boolean {
18
+
return existsSync(join(MODELS_DIR, file));
19
+
}
20
+
21
+
export function listModels(): void {
22
+
const activeChatId = getActiveChatModel().id;
23
+
const activeAutoId = getActiveAutocompleteModel().id;
24
+
25
+
console.log(`\n${BOLD}Chat models:${RESET}`);
26
+
for (const m of getChatModels()) {
27
+
const active = m.id === activeChatId ? ` ${GREEN}<- active${RESET}` : "";
28
+
const downloaded = isDownloaded(m.file) ? "" : ` ${DIM}(not downloaded)${RESET}`;
29
+
console.log(
30
+
` ${BOLD}${m.id}${RESET} ${m.name} ~${m.sizeGB}GB min ${m.minRAMGB}GB RAM${active}${downloaded}`,
31
+
);
32
+
}
33
+
34
+
console.log(`\n${BOLD}Autocomplete models:${RESET}`);
35
+
for (const m of getAutocompleteModels()) {
36
+
const active = m.id === activeAutoId ? ` ${GREEN}<- active${RESET}` : "";
37
+
const downloaded = isDownloaded(m.file) ? "" : ` ${DIM}(not downloaded)${RESET}`;
38
+
console.log(
39
+
` ${BOLD}${m.id}${RESET} ${m.name} ~${m.sizeGB}GB min ${m.minRAMGB}GB RAM${active}${downloaded}`,
40
+
);
41
+
}
42
+
console.log("");
43
+
}
44
+
45
+
export async function setChatModel(id: string): Promise<void> {
46
+
const model = getModelById(id);
47
+
if (!model) {
48
+
err(`Unknown model: ${id}\nAvailable: ${MODELS.map((m) => m.id).join(", ")}`);
49
+
}
50
+
if (model.role !== "chat") {
51
+
err(`${id} is an ${model.role} model, not a chat model.`);
52
+
}
53
+
54
+
const config = loadConfig();
55
+
config.chatModel = id;
56
+
saveConfig(config);
57
+
log(`Chat model set to ${model.name}`);
58
+
59
+
// Download if needed
60
+
const dest = join(MODELS_DIR, model.file);
61
+
if (!existsSync(dest)) {
62
+
mkdirSync(MODELS_DIR, { recursive: true });
63
+
log(`Downloading ${model.name} (${model.sizeGB}GB)...`);
64
+
runPassthrough(`curl -L --progress-bar -o "${dest}" "${model.url}"`);
65
+
log(`Downloaded: ${model.file}`);
66
+
}
67
+
68
+
// Regenerate scripts and configs
69
+
await createLauncherScripts();
70
+
await writeTuiConfig();
71
+
log("Launcher scripts and configs regenerated.");
72
+
log("Run llama-stop && llama-start to use the new model.");
73
+
}
74
+
75
+
export async function setAutocompleteModel(id: string): Promise<void> {
76
+
const model = getModelById(id);
77
+
if (!model) {
78
+
err(`Unknown model: ${id}\nAvailable: ${MODELS.map((m) => m.id).join(", ")}`);
79
+
}
80
+
if (model.role !== "autocomplete") {
81
+
err(`${id} is a ${model.role} model, not an autocomplete model.`);
82
+
}
83
+
84
+
const config = loadConfig();
85
+
config.autocompleteModel = id;
86
+
saveConfig(config);
87
+
log(`Autocomplete model set to ${model.name}`);
88
+
89
+
// Download if needed
90
+
const dest = join(MODELS_DIR, model.file);
91
+
if (!existsSync(dest)) {
92
+
mkdirSync(MODELS_DIR, { recursive: true });
93
+
log(`Downloading ${model.name} (${model.sizeGB}GB)...`);
94
+
runPassthrough(`curl -L --progress-bar -o "${dest}" "${model.url}"`);
95
+
log(`Downloaded: ${model.file}`);
96
+
}
97
+
98
+
await createLauncherScripts();
99
+
await writeTuiConfig();
100
+
log("Launcher scripts and configs regenerated.");
101
+
log("Run llama-stop && llama-start to use the new model.");
102
+
}
+48
src/commands/pipe.ts
+48
src/commands/pipe.ts
···
1
+
import { CHAT_PORT } from "../config.js";
2
+
import { err } from "../log.js";
3
+
4
+
export async function runPipe(prompt: string): Promise<void> {
5
+
// Read stdin
6
+
const chunks: Buffer[] = [];
7
+
for await (const chunk of process.stdin) {
8
+
chunks.push(chunk as Buffer);
9
+
}
10
+
const input = Buffer.concat(chunks).toString("utf-8");
11
+
12
+
const body = JSON.stringify({
13
+
model: "qwen",
14
+
messages: [
15
+
{
16
+
role: "system",
17
+
content:
18
+
"You are an expert programmer. Output only code, no explanations.",
19
+
},
20
+
{ role: "user", content: `${prompt}\n\n\`\`\`\n${input}\n\`\`\`` },
21
+
],
22
+
stream: false,
23
+
});
24
+
25
+
let res: Response;
26
+
try {
27
+
res = await fetch(
28
+
`http://127.0.0.1:${CHAT_PORT}/v1/chat/completions`,
29
+
{
30
+
method: "POST",
31
+
headers: { "Content-Type": "application/json" },
32
+
body,
33
+
},
34
+
);
35
+
} catch {
36
+
err(`Chat server not running on port ${CHAT_PORT}.\nStart it with: localcode start`);
37
+
}
38
+
39
+
if (!res!.ok) {
40
+
err(`Server returned ${res!.status}`);
41
+
}
42
+
43
+
const data = (await res!.json()) as {
44
+
choices?: { message?: { content?: string } }[];
45
+
};
46
+
const content = data.choices?.[0]?.message?.content ?? "";
47
+
process.stdout.write(content + "\n");
48
+
}
+90
src/commands/run.ts
+90
src/commands/run.ts
···
1
+
import { spawn, execSync } from "node:child_process";
2
+
import { existsSync } from "node:fs";
3
+
import { join } from "node:path";
4
+
import { CHAT_PORT, LAUNCH_DIR, AIDER_ENV_FILE } from "../config.js";
5
+
import { getActiveTui } from "../runtime-config.js";
6
+
import { log, warn } from "../log.js";
7
+
8
+
async function waitForHealth(
9
+
port: number,
10
+
timeoutSec: number,
11
+
): Promise<boolean> {
12
+
for (let i = 0; i < timeoutSec; i++) {
13
+
try {
14
+
const res = await fetch(`http://127.0.0.1:${port}/health`);
15
+
if (res.ok) return true;
16
+
} catch {
17
+
// not ready yet
18
+
}
19
+
await new Promise((r) => setTimeout(r, 2000));
20
+
process.stdout.write(".");
21
+
}
22
+
return false;
23
+
}
24
+
25
+
async function ensureServer(): Promise<void> {
26
+
try {
27
+
const res = await fetch(`http://127.0.0.1:${CHAT_PORT}/health`);
28
+
if (res.ok) return;
29
+
} catch {
30
+
// not running
31
+
}
32
+
33
+
const serverScript = join(LAUNCH_DIR, "llama-chat-server");
34
+
if (!existsSync(serverScript)) {
35
+
warn("Server script not found. Run: localcode setup");
36
+
process.exit(1);
37
+
}
38
+
39
+
log("Starting llama.cpp chat server...");
40
+
const child = spawn(serverScript, [], {
41
+
stdio: "ignore",
42
+
detached: true,
43
+
});
44
+
child.unref();
45
+
46
+
process.stdout.write("Waiting for model to load");
47
+
const ready = await waitForHealth(CHAT_PORT, 120);
48
+
if (ready) {
49
+
console.log(" ready!");
50
+
} else {
51
+
console.log(" timed out.");
52
+
warn("Server may still be loading. Check /tmp/llama-chat.log");
53
+
}
54
+
}
55
+
56
+
function ensureGit(): void {
57
+
if (!existsSync(".git")) {
58
+
log("Initializing git repo...");
59
+
execSync("git init", { stdio: "inherit" });
60
+
execSync("git add -A", { stdio: "inherit" });
61
+
execSync('git commit -m "Initial commit (before AI edits)" --allow-empty', {
62
+
stdio: "inherit",
63
+
});
64
+
}
65
+
}
66
+
67
+
export async function runDefault(args: string[]): Promise<void> {
68
+
await ensureServer();
69
+
ensureGit();
70
+
71
+
const tui = getActiveTui();
72
+
73
+
// Set up env for aider
74
+
if (tui.id === "aider" && existsSync(AIDER_ENV_FILE)) {
75
+
const { readFileSync } = await import("node:fs");
76
+
const env = readFileSync(AIDER_ENV_FILE, "utf-8");
77
+
for (const line of env.split("\n")) {
78
+
if (line && !line.startsWith("#")) {
79
+
const eq = line.indexOf("=");
80
+
if (eq > 0) {
81
+
process.env[line.slice(0, eq)] = line.slice(eq + 1);
82
+
}
83
+
}
84
+
}
85
+
}
86
+
87
+
log(`Launching ${tui.name}...`);
88
+
const child = spawn(tui.checkCmd, args, { stdio: "inherit" });
89
+
child.on("exit", (code) => process.exit(code ?? 0));
90
+
}
+96
src/commands/server.ts
+96
src/commands/server.ts
···
1
+
import { spawn, execSync } from "node:child_process";
2
+
import { existsSync } from "node:fs";
3
+
import { join } from "node:path";
4
+
import { CHAT_PORT, AUTOCOMPLETE_PORT, LAUNCH_DIR } from "../config.js";
5
+
import {
6
+
getActiveChatModel,
7
+
getActiveAutocompleteModel,
8
+
} from "../runtime-config.js";
9
+
import { log, warn, err } from "../log.js";
10
+
11
+
async function waitForHealth(port: number, label: string): Promise<boolean> {
12
+
process.stdout.write(` Waiting for ${label}`);
13
+
for (let i = 0; i < 60; i++) {
14
+
try {
15
+
const res = await fetch(`http://127.0.0.1:${port}/health`);
16
+
if (res.ok) {
17
+
console.log(" ready!");
18
+
return true;
19
+
}
20
+
} catch {
21
+
// not ready
22
+
}
23
+
process.stdout.write(".");
24
+
await new Promise((r) => setTimeout(r, 2000));
25
+
}
26
+
console.log(" timed out.");
27
+
return false;
28
+
}
29
+
30
+
export async function startServers(): Promise<void> {
31
+
const chatModel = getActiveChatModel();
32
+
const autoModel = getActiveAutocompleteModel();
33
+
34
+
const chatScript = join(LAUNCH_DIR, "llama-chat-server");
35
+
const autoScript = join(LAUNCH_DIR, "llama-complete-server");
36
+
37
+
if (!existsSync(chatScript)) {
38
+
err("Server scripts not found. Run: localcode setup");
39
+
}
40
+
41
+
log(`Starting servers...`);
42
+
43
+
// Chat server
44
+
let chatAlready = false;
45
+
try {
46
+
const res = await fetch(`http://127.0.0.1:${CHAT_PORT}/health`);
47
+
chatAlready = res.ok;
48
+
} catch {
49
+
// not running
50
+
}
51
+
52
+
if (chatAlready) {
53
+
log(`Chat server already running on :${CHAT_PORT}`);
54
+
} else {
55
+
log(`Chat: ${chatModel.name} on :${CHAT_PORT}`);
56
+
const c = spawn(chatScript, [], {
57
+
stdio: ["ignore", "ignore", "ignore"],
58
+
detached: true,
59
+
});
60
+
c.unref();
61
+
await waitForHealth(CHAT_PORT, "chat model");
62
+
}
63
+
64
+
// Autocomplete server
65
+
let autoAlready = false;
66
+
try {
67
+
const res = await fetch(`http://127.0.0.1:${AUTOCOMPLETE_PORT}/health`);
68
+
autoAlready = res.ok;
69
+
} catch {
70
+
// not running
71
+
}
72
+
73
+
if (autoAlready) {
74
+
log(`Autocomplete server already running on :${AUTOCOMPLETE_PORT}`);
75
+
} else {
76
+
log(`Autocomplete: ${autoModel.name} on :${AUTOCOMPLETE_PORT}`);
77
+
const c = spawn(autoScript, [], {
78
+
stdio: ["ignore", "ignore", "ignore"],
79
+
detached: true,
80
+
});
81
+
c.unref();
82
+
await waitForHealth(AUTOCOMPLETE_PORT, "autocomplete model");
83
+
}
84
+
85
+
console.log("");
86
+
log("Servers running. Logs: /tmp/llama-chat.log, /tmp/llama-complete.log");
87
+
}
88
+
89
+
export function stopServers(): void {
90
+
try {
91
+
execSync('pkill -f "llama-server"', { stdio: "ignore" });
92
+
log("Servers stopped.");
93
+
} catch {
94
+
warn("No servers running.");
95
+
}
96
+
}
+91
src/commands/setup.ts
+91
src/commands/setup.ts
···
1
+
import { checkPreflight } from "../steps/preflight.js";
2
+
import { installHomebrew } from "../steps/homebrew.js";
3
+
import { installLlama } from "../steps/llama.js";
4
+
import { downloadModels } from "../steps/models.js";
5
+
import { installTools } from "../steps/tools.js";
6
+
import { createLauncherScripts } from "../steps/scripts.js";
7
+
import { writeTuiConfig } from "../steps/aider-config.js";
8
+
import { addToPath } from "../steps/shell-path.js";
9
+
import { MODELS_DIR, AIDER_CONFIG_FILE, AIDER_CONFIG_DIR } from "../config.js";
10
+
import {
11
+
getActiveChatModel,
12
+
getActiveAutocompleteModel,
13
+
getActiveTui,
14
+
} from "../runtime-config.js";
15
+
16
+
const BOLD = "\x1b[1m";
17
+
const GREEN = "\x1b[0;32m";
18
+
const RESET = "\x1b[0m";
19
+
20
+
function printSummary(): void {
21
+
const chatModel = getActiveChatModel();
22
+
const autocompleteModel = getActiveAutocompleteModel();
23
+
const tui = getActiveTui();
24
+
const shellRC =
25
+
process.env.SHELL?.endsWith("/zsh") !== false
26
+
? "~/.zshrc"
27
+
: process.env.SHELL?.endsWith("/bash")
28
+
? "~/.bashrc"
29
+
: "~/.profile";
30
+
31
+
console.log("");
32
+
console.log(
33
+
`${GREEN}${BOLD}═══════════════════════════════════════════════════${RESET}`,
34
+
);
35
+
console.log(`${GREEN}${BOLD} Setup complete!${RESET}`);
36
+
console.log(
37
+
`${GREEN}${BOLD}═══════════════════════════════════════════════════${RESET}`,
38
+
);
39
+
console.log("");
40
+
console.log(` ${BOLD}Models downloaded to:${RESET} ${MODELS_DIR}`);
41
+
console.log(
42
+
` Chat: ${chatModel.name} (${chatModel.file}, ~${chatModel.sizeGB}GB)`,
43
+
);
44
+
console.log(
45
+
` Autocomplete: ${autocompleteModel.name} (${autocompleteModel.file}, ~${autocompleteModel.sizeGB}GB)`,
46
+
);
47
+
console.log("");
48
+
console.log(` ${BOLD}Active TUI:${RESET} ${tui.name}`);
49
+
console.log("");
50
+
console.log(` ${BOLD}Commands available${RESET} (restart your shell first):`);
51
+
console.log("");
52
+
console.log(` ${BOLD}llama-start${RESET} Start both llama.cpp servers`);
53
+
console.log(` ${BOLD}llama-stop${RESET} Stop all llama.cpp servers`);
54
+
console.log("");
55
+
console.log(
56
+
` ${BOLD}ai-code${RESET} [dir] Full coding agent (auto-starts server)`,
57
+
);
58
+
console.log(
59
+
` ${BOLD}ai-ask${RESET} "question" Quick coding Q&A, no file edits`,
60
+
);
61
+
console.log(
62
+
` ${BOLD}ai-pipe${RESET} "prompt" Pipe code through the model`,
63
+
);
64
+
console.log("");
65
+
console.log(` ${BOLD}Config:${RESET} ${AIDER_CONFIG_FILE}`);
66
+
console.log(` ${BOLD}API env:${RESET} ${AIDER_CONFIG_DIR}/.env`);
67
+
console.log(
68
+
` ${BOLD}Server logs:${RESET} /tmp/llama-chat.log, /tmp/llama-complete.log`,
69
+
);
70
+
console.log("");
71
+
console.log(
72
+
` Run ${BOLD}source ${shellRC}${RESET} or open a new terminal to get started.`,
73
+
);
74
+
console.log("");
75
+
}
76
+
77
+
export async function runSetup(): Promise<void> {
78
+
console.log(
79
+
`\n${BOLD}Local AI Coding Environment Installer (llama.cpp)${RESET}\n`,
80
+
);
81
+
82
+
checkPreflight();
83
+
installHomebrew();
84
+
installLlama();
85
+
downloadModels();
86
+
installTools();
87
+
await createLauncherScripts();
88
+
await writeTuiConfig();
89
+
addToPath();
90
+
printSummary();
91
+
}
+45
src/commands/status.ts
+45
src/commands/status.ts
···
1
+
import { CHAT_PORT, AUTOCOMPLETE_PORT, LAUNCH_DIR } from "../config.js";
2
+
import {
3
+
getActiveChatModel,
4
+
getActiveAutocompleteModel,
5
+
getActiveTui,
6
+
} from "../runtime-config.js";
7
+
8
+
const BOLD = "\x1b[1m";
9
+
const GREEN = "\x1b[0;32m";
10
+
const RED = "\x1b[0;31m";
11
+
const DIM = "\x1b[2m";
12
+
const RESET = "\x1b[0m";
13
+
14
+
async function checkHealth(port: number): Promise<boolean> {
15
+
try {
16
+
const res = await fetch(`http://127.0.0.1:${port}/health`);
17
+
return res.ok;
18
+
} catch {
19
+
return false;
20
+
}
21
+
}
22
+
23
+
export async function showStatus(): Promise<void> {
24
+
const chatModel = getActiveChatModel();
25
+
const autoModel = getActiveAutocompleteModel();
26
+
const tui = getActiveTui();
27
+
28
+
const chatOk = await checkHealth(CHAT_PORT);
29
+
const autoOk = await checkHealth(AUTOCOMPLETE_PORT);
30
+
31
+
const on = `${GREEN}running${RESET}`;
32
+
const off = `${RED}stopped${RESET}`;
33
+
34
+
console.log(`
35
+
${BOLD}localcode${RESET} — current configuration
36
+
37
+
${BOLD}Chat model:${RESET} ${chatModel.name} ${DIM}(${chatModel.id})${RESET}
38
+
${BOLD}Autocomplete model:${RESET} ${autoModel.name} ${DIM}(${autoModel.id})${RESET}
39
+
${BOLD}Active TUI:${RESET} ${tui.name} ${DIM}(${tui.id})${RESET}
40
+
41
+
${BOLD}Chat server:${RESET} :${CHAT_PORT} ${chatOk ? on : off}
42
+
${BOLD}Autocomplete:${RESET} :${AUTOCOMPLETE_PORT} ${autoOk ? on : off}
43
+
${BOLD}Scripts:${RESET} ${LAUNCH_DIR}
44
+
`);
45
+
}
+50
src/commands/tuis.ts
+50
src/commands/tuis.ts
···
1
+
import { TUIS, getTuiById } from "../registry/tuis.js";
2
+
import { loadConfig, saveConfig, getActiveTui } from "../runtime-config.js";
3
+
import { createLauncherScripts } from "../steps/scripts.js";
4
+
import { writeTuiConfig } from "../steps/aider-config.js";
5
+
import { commandExists, runPassthrough } from "../util.js";
6
+
import { log, err } from "../log.js";
7
+
8
+
const BOLD = "\x1b[1m";
9
+
const GREEN = "\x1b[0;32m";
10
+
const DIM = "\x1b[2m";
11
+
const RESET = "\x1b[0m";
12
+
13
+
export function listTuis(): void {
14
+
const activeId = getActiveTui().id;
15
+
16
+
console.log(`\n${BOLD}Available TUIs:${RESET}`);
17
+
for (const t of TUIS) {
18
+
const active = t.id === activeId ? ` ${GREEN}<- active${RESET}` : "";
19
+
const installed = commandExists(t.checkCmd)
20
+
? ""
21
+
: ` ${DIM}(not installed)${RESET}`;
22
+
console.log(` ${BOLD}${t.id}${RESET} ${t.name}${active}${installed}`);
23
+
}
24
+
console.log("");
25
+
}
26
+
27
+
export async function setTui(id: string): Promise<void> {
28
+
const tui = getTuiById(id);
29
+
if (!tui) {
30
+
err(
31
+
`Unknown TUI: ${id}\nAvailable: ${TUIS.map((t) => t.id).join(", ")}`,
32
+
);
33
+
}
34
+
35
+
// Install if needed
36
+
if (!commandExists(tui.checkCmd)) {
37
+
log(`Installing ${tui.name}...`);
38
+
runPassthrough(tui.installCmd);
39
+
}
40
+
41
+
const config = loadConfig();
42
+
config.tui = id;
43
+
saveConfig(config);
44
+
log(`Active TUI set to ${tui.name}`);
45
+
46
+
// Regenerate scripts and configs
47
+
await createLauncherScripts();
48
+
await writeTuiConfig();
49
+
log("Launcher scripts and configs regenerated.");
50
+
}
+14
src/config.ts
+14
src/config.ts
···
1
+
import { homedir } from "node:os";
2
+
import { join } from "node:path";
3
+
4
+
export const MODELS_DIR = join(homedir(), ".local/share/llama-models");
5
+
export const CHAT_PORT = 8080;
6
+
export const AUTOCOMPLETE_PORT = 8081;
7
+
export const LAUNCH_DIR = join(homedir(), ".local/bin");
8
+
export const AIDER_CONFIG_DIR = join(homedir(), ".aider");
9
+
export const AIDER_CONFIG_FILE = join(AIDER_CONFIG_DIR, "aider.conf.yml");
10
+
export const AIDER_ENV_FILE = join(AIDER_CONFIG_DIR, ".env");
11
+
export const OPENCODE_CONFIG_DIR = join(homedir(), ".config", "opencode");
12
+
export const OPENCODE_CONFIG_FILE = join(OPENCODE_CONFIG_DIR, "opencode.json");
13
+
export const PI_CONFIG_DIR = join(homedir(), ".pi", "agent");
14
+
export const PI_MODELS_FILE = join(PI_CONFIG_DIR, "models.json");
+18
src/log.ts
+18
src/log.ts
···
1
+
const BOLD = "\x1b[1m";
2
+
const GREEN = "\x1b[0;32m";
3
+
const YELLOW = "\x1b[1;33m";
4
+
const RED = "\x1b[0;31m";
5
+
const RESET = "\x1b[0m";
6
+
7
+
export function log(msg: string): void {
8
+
console.log(`${GREEN}${BOLD}[✓]${RESET} ${msg}`);
9
+
}
10
+
11
+
export function warn(msg: string): void {
12
+
console.log(`${YELLOW}${BOLD}[!]${RESET} ${msg}`);
13
+
}
14
+
15
+
export function err(msg: string): never {
16
+
console.error(`${RED}${BOLD}[✗]${RESET} ${msg}`);
17
+
process.exit(1);
18
+
}
+144
src/main.ts
+144
src/main.ts
···
1
+
import { runSetup } from "./commands/setup.js";
2
+
import { listModels, setChatModel, setAutocompleteModel } from "./commands/models.js";
3
+
import { listTuis, setTui } from "./commands/tuis.js";
4
+
import { runBench } from "./commands/bench.js";
5
+
import { runDefault } from "./commands/run.js";
6
+
import { showStatus } from "./commands/status.js";
7
+
import { startServers, stopServers } from "./commands/server.js";
8
+
import { runPipe } from "./commands/pipe.js";
9
+
10
+
const BOLD = "\x1b[1m";
11
+
const DIM = "\x1b[2m";
12
+
const RESET = "\x1b[0m";
13
+
14
+
function printUsage(): void {
15
+
console.log(`
16
+
${BOLD}localcode${RESET} — local AI coding environment
17
+
18
+
${BOLD}Usage:${RESET}
19
+
localcode ${DIM}[flags...]${RESET} Launch active TUI in current directory
20
+
localcode status Show current config and server health
21
+
22
+
${BOLD}Server:${RESET}
23
+
localcode start Start chat + autocomplete servers
24
+
localcode stop Stop all servers
25
+
26
+
${BOLD}Models:${RESET}
27
+
localcode models List available models
28
+
localcode models set-chat <id> Switch the chat model
29
+
localcode models set-auto <id> Switch the autocomplete model
30
+
31
+
${BOLD}TUIs:${RESET}
32
+
localcode tuis List available TUIs
33
+
localcode tuis set <id> Switch the active TUI
34
+
35
+
${BOLD}Benchmark:${RESET}
36
+
localcode bench Benchmark the running chat model
37
+
localcode bench history Show past benchmark results
38
+
39
+
${BOLD}Other:${RESET}
40
+
localcode pipe "prompt" Pipe stdin through the model
41
+
localcode setup Full install (models, tools, scripts)
42
+
`);
43
+
}
44
+
45
+
async function main(): Promise<void> {
46
+
const cmd = process.argv[2];
47
+
48
+
switch (cmd) {
49
+
case undefined:
50
+
await runDefault(process.argv.slice(3));
51
+
break;
52
+
53
+
case "status":
54
+
await showStatus();
55
+
break;
56
+
57
+
case "start":
58
+
await startServers();
59
+
break;
60
+
61
+
case "stop":
62
+
stopServers();
63
+
break;
64
+
65
+
case "models": {
66
+
const sub = process.argv[3];
67
+
if (!sub) {
68
+
listModels();
69
+
} else if (sub === "set-chat") {
70
+
const id = process.argv[4];
71
+
if (!id) {
72
+
console.error("Usage: localcode models set-chat <model-id>");
73
+
process.exit(1);
74
+
}
75
+
await setChatModel(id);
76
+
} else if (sub === "set-auto" || sub === "set-autocomplete") {
77
+
const id = process.argv[4];
78
+
if (!id) {
79
+
console.error("Usage: localcode models set-auto <model-id>");
80
+
process.exit(1);
81
+
}
82
+
await setAutocompleteModel(id);
83
+
} else {
84
+
console.error(`Unknown: localcode models ${sub}`);
85
+
process.exit(1);
86
+
}
87
+
break;
88
+
}
89
+
90
+
case "tuis": {
91
+
const sub = process.argv[3];
92
+
if (!sub) {
93
+
listTuis();
94
+
} else if (sub === "set") {
95
+
const id = process.argv[4];
96
+
if (!id) {
97
+
console.error("Usage: localcode tuis set <tui-id>");
98
+
process.exit(1);
99
+
}
100
+
await setTui(id);
101
+
} else {
102
+
console.error(`Unknown: localcode tuis ${sub}`);
103
+
process.exit(1);
104
+
}
105
+
break;
106
+
}
107
+
108
+
case "bench": {
109
+
const sub = process.argv[3];
110
+
if (sub === "history") {
111
+
await runBench(["--history"]);
112
+
} else {
113
+
await runBench(process.argv.slice(3));
114
+
}
115
+
break;
116
+
}
117
+
118
+
case "pipe": {
119
+
const prompt = process.argv[3] ?? "Improve this code";
120
+
await runPipe(prompt);
121
+
break;
122
+
}
123
+
124
+
case "setup":
125
+
await runSetup();
126
+
break;
127
+
128
+
case "help":
129
+
case "--help":
130
+
case "-h":
131
+
printUsage();
132
+
break;
133
+
134
+
default:
135
+
console.error(`Unknown command: ${cmd}`);
136
+
printUsage();
137
+
process.exit(1);
138
+
}
139
+
}
140
+
141
+
main().catch((e: unknown) => {
142
+
console.error(e);
143
+
process.exit(1);
144
+
});
+55
src/registry/models.ts
+55
src/registry/models.ts
···
1
+
export interface ModelDef {
2
+
id: string;
3
+
name: string;
4
+
role: "chat" | "autocomplete";
5
+
file: string;
6
+
url: string;
7
+
sizeGB: number;
8
+
ctxSize: number;
9
+
minRAMGB: number;
10
+
}
11
+
12
+
export const MODELS: ModelDef[] = [
13
+
{
14
+
id: "qwen-32b-chat",
15
+
name: "Qwen 2.5 Coder 32B",
16
+
role: "chat",
17
+
file: "qwen2.5-coder-32b-instruct-q4_k_m.gguf",
18
+
url: "https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct-GGUF/resolve/main/qwen2.5-coder-32b-instruct-q4_k_m.gguf",
19
+
sizeGB: 20,
20
+
ctxSize: 16384,
21
+
minRAMGB: 32,
22
+
},
23
+
{
24
+
id: "qwen-14b-chat",
25
+
name: "Qwen 2.5 Coder 14B",
26
+
role: "chat",
27
+
file: "qwen2.5-coder-14b-instruct-q4_k_m.gguf",
28
+
url: "https://huggingface.co/Qwen/Qwen2.5-Coder-14B-Instruct-GGUF/resolve/main/qwen2.5-coder-14b-instruct-q4_k_m.gguf",
29
+
sizeGB: 9,
30
+
ctxSize: 16384,
31
+
minRAMGB: 16,
32
+
},
33
+
{
34
+
id: "qwen-1.5b-autocomplete",
35
+
name: "Qwen 2.5 Coder 1.5B",
36
+
role: "autocomplete",
37
+
file: "qwen2.5-coder-1.5b-instruct-q4_k_m.gguf",
38
+
url: "https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF/resolve/main/qwen2.5-coder-1.5b-instruct-q4_k_m.gguf",
39
+
sizeGB: 1.2,
40
+
ctxSize: 4096,
41
+
minRAMGB: 8,
42
+
},
43
+
];
44
+
45
+
export function getModelById(id: string): ModelDef | undefined {
46
+
return MODELS.find((m) => m.id === id);
47
+
}
48
+
49
+
export function getChatModels(): ModelDef[] {
50
+
return MODELS.filter((m) => m.role === "chat");
51
+
}
52
+
53
+
export function getAutocompleteModels(): ModelDef[] {
54
+
return MODELS.filter((m) => m.role === "autocomplete");
55
+
}
+35
src/registry/tuis.ts
+35
src/registry/tuis.ts
···
1
+
export interface TuiDef {
2
+
id: string;
3
+
name: string;
4
+
installCmd: string;
5
+
checkCmd: string;
6
+
launchArgs: string;
7
+
}
8
+
9
+
export const TUIS: TuiDef[] = [
10
+
{
11
+
id: "aider",
12
+
name: "Aider",
13
+
installCmd: "pipx install aider-chat",
14
+
checkCmd: "aider",
15
+
launchArgs: '"$@"',
16
+
},
17
+
{
18
+
id: "opencode",
19
+
name: "OpenCode",
20
+
installCmd: "npm install -g opencode-ai@latest",
21
+
checkCmd: "opencode",
22
+
launchArgs: '"$@"',
23
+
},
24
+
{
25
+
id: "pi",
26
+
name: "Pi",
27
+
installCmd: "npm install -g @mariozechner/pi-coding-agent",
28
+
checkCmd: "pi",
29
+
launchArgs: '"$@"',
30
+
},
31
+
];
32
+
33
+
export function getTuiById(id: string): TuiDef | undefined {
34
+
return TUIS.find((t) => t.id === id);
35
+
}
+54
src/runtime-config.ts
+54
src/runtime-config.ts
···
1
+
import { readFileSync, writeFileSync, mkdirSync } from "node:fs";
2
+
import { dirname, join } from "node:path";
3
+
import { homedir } from "node:os";
4
+
import { getModelById, MODELS } from "./registry/models.js";
5
+
import { getTuiById, TUIS } from "./registry/tuis.js";
6
+
import type { ModelDef } from "./registry/models.js";
7
+
import type { TuiDef } from "./registry/tuis.js";
8
+
9
+
export interface RuntimeConfig {
10
+
chatModel: string;
11
+
autocompleteModel: string;
12
+
tui: string;
13
+
}
14
+
15
+
const CONFIG_PATH = join(homedir(), ".config", "localcode", "config.json");
16
+
17
+
const DEFAULTS: RuntimeConfig = {
18
+
chatModel: "qwen-32b-chat",
19
+
autocompleteModel: "qwen-1.5b-autocomplete",
20
+
tui: "aider",
21
+
};
22
+
23
+
export function loadConfig(): RuntimeConfig {
24
+
try {
25
+
const raw = readFileSync(CONFIG_PATH, "utf-8");
26
+
const parsed = JSON.parse(raw) as Partial<RuntimeConfig>;
27
+
return { ...DEFAULTS, ...parsed };
28
+
} catch {
29
+
return { ...DEFAULTS };
30
+
}
31
+
}
32
+
33
+
export function saveConfig(config: RuntimeConfig): void {
34
+
mkdirSync(dirname(CONFIG_PATH), { recursive: true });
35
+
writeFileSync(CONFIG_PATH, JSON.stringify(config, null, 2) + "\n");
36
+
}
37
+
38
+
export function getActiveChatModel(): ModelDef {
39
+
const config = loadConfig();
40
+
return getModelById(config.chatModel) ?? MODELS[0]!;
41
+
}
42
+
43
+
export function getActiveAutocompleteModel(): ModelDef {
44
+
const config = loadConfig();
45
+
return (
46
+
getModelById(config.autocompleteModel) ??
47
+
MODELS.find((m) => m.role === "autocomplete")!
48
+
);
49
+
}
50
+
51
+
export function getActiveTui(): TuiDef {
52
+
const config = loadConfig();
53
+
return getTuiById(config.tui) ?? TUIS[0]!;
54
+
}
+36
src/steps/aider-config.ts
+36
src/steps/aider-config.ts
···
1
+
import { log } from "../log.js";
2
+
import { writeConfig } from "../util.js";
3
+
import {
4
+
AIDER_CONFIG_FILE,
5
+
AIDER_ENV_FILE,
6
+
OPENCODE_CONFIG_FILE,
7
+
PI_MODELS_FILE,
8
+
} from "../config.js";
9
+
import { getActiveChatModel, getActiveTui } from "../runtime-config.js";
10
+
import { aiderConfig, aiderEnv } from "../templates/aider.js";
11
+
import { opencodeConfig } from "../templates/opencode.js";
12
+
import { piModelsConfig } from "../templates/pi.js";
13
+
14
+
export async function writeTuiConfig(): Promise<void> {
15
+
const chatModel = getActiveChatModel();
16
+
const tui = getActiveTui();
17
+
18
+
// Always write all configs so switching TUIs doesn't require re-running setup
19
+
await writeConfig(AIDER_CONFIG_FILE, aiderConfig(chatModel.id));
20
+
await writeConfig(AIDER_ENV_FILE, aiderEnv());
21
+
log(`Aider config written to ${AIDER_CONFIG_FILE}`);
22
+
23
+
await writeConfig(
24
+
OPENCODE_CONFIG_FILE,
25
+
opencodeConfig(chatModel.id, chatModel.name),
26
+
);
27
+
log(`OpenCode config written to ${OPENCODE_CONFIG_FILE}`);
28
+
29
+
await writeConfig(
30
+
PI_MODELS_FILE,
31
+
piModelsConfig(chatModel.id, chatModel.name),
32
+
);
33
+
log(`Pi config written to ${PI_MODELS_FILE}`);
34
+
35
+
log(`Active TUI: ${tui.name}`);
36
+
}
+15
src/steps/homebrew.ts
+15
src/steps/homebrew.ts
···
1
+
import { log } from "../log.js";
2
+
import { commandExists, runPassthrough } from "../util.js";
3
+
4
+
export function installHomebrew(): void {
5
+
if (commandExists("brew")) {
6
+
log("Homebrew already installed.");
7
+
return;
8
+
}
9
+
10
+
log("Installing Homebrew...");
11
+
runPassthrough(
12
+
'/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"',
13
+
);
14
+
runPassthrough('eval "$(/opt/homebrew/bin/brew shellenv)"');
15
+
}
+23
src/steps/llama.ts
+23
src/steps/llama.ts
···
1
+
import { log, warn } from "../log.js";
2
+
import { commandExists, run, runPassthrough } from "../util.js";
3
+
4
+
export function installLlama(): void {
5
+
if (commandExists("llama-server")) {
6
+
log("llama.cpp already installed.");
7
+
} else {
8
+
log("Installing llama.cpp via Homebrew...");
9
+
runPassthrough("brew install llama.cpp");
10
+
}
11
+
12
+
// Verify Metal support
13
+
try {
14
+
const help = run("llama-server --help 2>&1", { silent: true });
15
+
if (/metal/i.test(help)) {
16
+
log("Metal (GPU) acceleration available.");
17
+
} else {
18
+
warn("Metal flag not detected — model will run on CPU only.");
19
+
}
20
+
} catch {
21
+
warn("Could not verify Metal support.");
22
+
}
23
+
}
+25
src/steps/models.ts
+25
src/steps/models.ts
···
1
+
import { existsSync, mkdirSync } from "node:fs";
2
+
import { join } from "node:path";
3
+
import { log } from "../log.js";
4
+
import { runPassthrough } from "../util.js";
5
+
import { MODELS_DIR } from "../config.js";
6
+
import { getActiveChatModel, getActiveAutocompleteModel } from "../runtime-config.js";
7
+
import type { ModelDef } from "../registry/models.js";
8
+
9
+
function downloadModel(model: ModelDef): void {
10
+
const dest = join(MODELS_DIR, model.file);
11
+
if (existsSync(dest)) {
12
+
log(`Model already downloaded: ${model.file}`);
13
+
return;
14
+
}
15
+
16
+
log(`Downloading ${model.name} (${model.sizeGB}GB, this may take a while)...`);
17
+
runPassthrough(`curl -L --progress-bar -o "${dest}" "${model.url}"`);
18
+
log(`Downloaded: ${model.file}`);
19
+
}
20
+
21
+
export function downloadModels(): void {
22
+
mkdirSync(MODELS_DIR, { recursive: true });
23
+
downloadModel(getActiveChatModel());
24
+
downloadModel(getActiveAutocompleteModel());
25
+
}
+19
src/steps/preflight.ts
+19
src/steps/preflight.ts
···
1
+
import { platform, arch, totalmem } from "node:os";
2
+
import { log, warn, err } from "../log.js";
3
+
4
+
export function checkPreflight(): void {
5
+
if (platform() !== "darwin") {
6
+
err("This script is for macOS only.");
7
+
}
8
+
9
+
if (arch() !== "arm64") {
10
+
warn("Not running on Apple Silicon — performance may vary.");
11
+
}
12
+
13
+
const memGB = Math.floor(totalmem() / 1073741824);
14
+
if (memGB < 32) {
15
+
warn(
16
+
`You have ${memGB}GB RAM. The 32B model needs ~20GB; you may experience swapping.`,
17
+
);
18
+
}
19
+
}
+32
src/steps/scripts.ts
+32
src/steps/scripts.ts
···
1
+
import { join } from "node:path";
2
+
import { log } from "../log.js";
3
+
import { writeExecutable } from "../util.js";
4
+
import { LAUNCH_DIR } from "../config.js";
5
+
import {
6
+
getActiveChatModel,
7
+
getActiveAutocompleteModel,
8
+
} from "../runtime-config.js";
9
+
import {
10
+
llamaChatServer,
11
+
llamaCompleteServer,
12
+
localcodeWrapper,
13
+
} from "../templates/scripts.js";
14
+
15
+
export async function createLauncherScripts(): Promise<void> {
16
+
const chatModel = getActiveChatModel();
17
+
const autocompleteModel = getActiveAutocompleteModel();
18
+
19
+
// Resolve the project directory (where dist/main.js lives)
20
+
const projectDir = join(import.meta.dirname, "../..");
21
+
22
+
const scripts: [string, string][] = [
23
+
["llama-chat-server", llamaChatServer(chatModel)],
24
+
["llama-complete-server", llamaCompleteServer(autocompleteModel)],
25
+
["localcode", localcodeWrapper(projectDir)],
26
+
];
27
+
28
+
for (const [name, content] of scripts) {
29
+
await writeExecutable(join(LAUNCH_DIR, name), content);
30
+
}
31
+
log(`Launcher scripts written to ${LAUNCH_DIR}`);
32
+
}
+34
src/steps/shell-path.ts
+34
src/steps/shell-path.ts
···
1
+
import { readFileSync, appendFileSync } from "node:fs";
2
+
import { homedir } from "node:os";
3
+
import { join } from "node:path";
4
+
import { log } from "../log.js";
5
+
6
+
export function addToPath(): void {
7
+
const shell = process.env.SHELL ?? "/bin/zsh";
8
+
let rcFile: string;
9
+
10
+
if (shell.endsWith("/zsh")) {
11
+
rcFile = join(homedir(), ".zshrc");
12
+
} else if (shell.endsWith("/bash")) {
13
+
rcFile = join(homedir(), ".bashrc");
14
+
} else {
15
+
rcFile = join(homedir(), ".profile");
16
+
}
17
+
18
+
let contents = "";
19
+
try {
20
+
contents = readFileSync(rcFile, "utf-8");
21
+
} catch {
22
+
// File doesn't exist yet, that's fine
23
+
}
24
+
25
+
if (contents.includes(".local/bin")) {
26
+
return;
27
+
}
28
+
29
+
appendFileSync(
30
+
rcFile,
31
+
'\n# Local AI coding tools\nexport PATH="$HOME/.local/bin:$PATH"\n',
32
+
);
33
+
log(`Added ~/.local/bin to PATH in ${rcFile}`);
34
+
}
+27
src/steps/tools.ts
+27
src/steps/tools.ts
···
1
+
import { log } from "../log.js";
2
+
import { commandExists, runPassthrough } from "../util.js";
3
+
import { getActiveTui } from "../runtime-config.js";
4
+
5
+
function brewInstallIfMissing(cmd: string, pkg?: string): void {
6
+
if (commandExists(cmd)) {
7
+
log(`${pkg ?? cmd} already installed.`);
8
+
} else {
9
+
log(`Installing ${pkg ?? cmd}...`);
10
+
runPassthrough(`brew install ${pkg ?? cmd}`);
11
+
}
12
+
}
13
+
14
+
export function installTools(): void {
15
+
brewInstallIfMissing("jq");
16
+
brewInstallIfMissing("python3", "python@3.12");
17
+
brewInstallIfMissing("pipx");
18
+
19
+
const tui = getActiveTui();
20
+
21
+
if (!commandExists(tui.checkCmd)) {
22
+
log(`Installing ${tui.name}...`);
23
+
runPassthrough(tui.installCmd);
24
+
} else {
25
+
log(`${tui.name} already installed.`);
26
+
}
27
+
}
+39
src/templates/aider.ts
+39
src/templates/aider.ts
···
1
+
export function aiderConfig(modelName: string): string {
2
+
return `# =============================================================================
3
+
# Aider Configuration — ${modelName} via llama.cpp
4
+
# =============================================================================
5
+
6
+
# Point Aider at llama.cpp's OpenAI-compatible endpoint
7
+
# The model name can be anything — llama.cpp ignores it and uses the loaded model
8
+
model: openai/${modelName}
9
+
10
+
# Architect mode for better code planning
11
+
architect: true
12
+
editor-model: openai/${modelName}
13
+
14
+
# Git integration
15
+
auto-commits: true
16
+
dirty-commits: true
17
+
attribute-author: false
18
+
attribute-committer: false
19
+
20
+
# UI preferences
21
+
pretty: true
22
+
stream: true
23
+
dark-mode: true
24
+
25
+
# Code style
26
+
code-theme: monokai
27
+
show-diffs: true
28
+
29
+
# Disable analytics
30
+
analytics-disable: true
31
+
`;
32
+
}
33
+
34
+
export function aiderEnv(): string {
35
+
return `# llama.cpp serves an OpenAI-compatible API — no real key needed
36
+
OPENAI_API_KEY=sk-not-needed
37
+
OPENAI_API_BASE=http://127.0.0.1:8080/v1
38
+
`;
39
+
}
+28
src/templates/opencode.ts
+28
src/templates/opencode.ts
···
1
+
import { CHAT_PORT } from "../config.js";
2
+
3
+
export function opencodeConfig(modelId: string, modelName: string): string {
4
+
return JSON.stringify(
5
+
{
6
+
$schema: "https://opencode.ai/config.json",
7
+
model: `llama-cpp/${modelId}`,
8
+
provider: {
9
+
"llama-cpp": {
10
+
npm: "@ai-sdk/openai-compatible",
11
+
name: "llama.cpp (local)",
12
+
options: {
13
+
baseURL: `http://127.0.0.1:${CHAT_PORT}/v1`,
14
+
apiKey: "not-needed",
15
+
},
16
+
models: {
17
+
[modelId]: {
18
+
name: modelName,
19
+
tools: true,
20
+
},
21
+
},
22
+
},
23
+
},
24
+
},
25
+
null,
26
+
2,
27
+
) + "\n";
28
+
}
+23
src/templates/pi.ts
+23
src/templates/pi.ts
···
1
+
import { CHAT_PORT } from "../config.js";
2
+
3
+
export function piModelsConfig(modelId: string, modelName: string): string {
4
+
return JSON.stringify(
5
+
{
6
+
providers: {
7
+
"llama-cpp": {
8
+
baseUrl: `http://127.0.0.1:${CHAT_PORT}/v1`,
9
+
api: "openai-completions",
10
+
apiKey: "not-needed",
11
+
models: [
12
+
{
13
+
id: modelId,
14
+
name: modelName,
15
+
},
16
+
],
17
+
},
18
+
},
19
+
},
20
+
null,
21
+
2,
22
+
) + "\n";
23
+
}
+50
src/templates/scripts.ts
+50
src/templates/scripts.ts
···
1
+
import { MODELS_DIR, CHAT_PORT, AUTOCOMPLETE_PORT } from "../config.js";
2
+
import type { ModelDef } from "../registry/models.js";
3
+
4
+
// Use this to emit a literal $ in template literals without triggering interpolation.
5
+
const D = "$";
6
+
7
+
export function llamaChatServer(model: ModelDef): string {
8
+
return `#!/usr/bin/env bash
9
+
# Start llama.cpp server with ${model.name} for chat
10
+
# Exposed as OpenAI-compatible API on port ${CHAT_PORT}
11
+
12
+
MODEL="${MODELS_DIR}/${model.file}"
13
+
14
+
exec llama-server \\
15
+
--model "${D}MODEL" \\
16
+
--port ${CHAT_PORT} \\
17
+
--host 127.0.0.1 \\
18
+
--ctx-size ${model.ctxSize} \\
19
+
--n-gpu-layers 99 \\
20
+
--threads ${D}(sysctl -n hw.perflevel0.logicalcpu 2>/dev/null || echo 4) \\
21
+
--mlock \\
22
+
"${D}@"
23
+
`;
24
+
}
25
+
26
+
export function llamaCompleteServer(model: ModelDef): string {
27
+
return `#!/usr/bin/env bash
28
+
# Start llama.cpp server with ${model.name} for autocomplete
29
+
# Exposed as OpenAI-compatible API on port ${AUTOCOMPLETE_PORT}
30
+
31
+
MODEL="${MODELS_DIR}/${model.file}"
32
+
33
+
exec llama-server \\
34
+
--model "${D}MODEL" \\
35
+
--port ${AUTOCOMPLETE_PORT} \\
36
+
--host 127.0.0.1 \\
37
+
--ctx-size ${model.ctxSize} \\
38
+
--n-gpu-layers 99 \\
39
+
--threads ${D}(sysctl -n hw.perflevel0.logicalcpu 2>/dev/null || echo 4) \\
40
+
--mlock \\
41
+
"${D}@"
42
+
`;
43
+
}
44
+
45
+
export function localcodeWrapper(projectDir: string): string {
46
+
return `#!/usr/bin/env bash
47
+
# localcode — local AI coding environment manager
48
+
exec node "${projectDir}/dist/main.js" "${D}@"
49
+
`;
50
+
}
+48
src/util.ts
+48
src/util.ts
···
1
+
import { execSync, execFileSync } from "node:child_process";
2
+
import { writeFile, mkdir } from "node:fs/promises";
3
+
import { dirname } from "node:path";
4
+
5
+
export function commandExists(name: string): boolean {
6
+
try {
7
+
execSync(`command -v ${name}`, { stdio: "ignore" });
8
+
return true;
9
+
} catch {
10
+
return false;
11
+
}
12
+
}
13
+
14
+
export function run(cmd: string, opts?: { silent?: boolean }): string {
15
+
const stdio = opts?.silent ? "pipe" : "inherit";
16
+
const result = execSync(cmd, {
17
+
stdio: [stdio, "pipe", stdio],
18
+
encoding: "utf-8",
19
+
});
20
+
return result?.trim() ?? "";
21
+
}
22
+
23
+
export function runPassthrough(cmd: string): void {
24
+
execSync(cmd, { stdio: "inherit" });
25
+
}
26
+
27
+
export function runShell(cmd: string): string {
28
+
return execFileSync("/bin/bash", ["-c", cmd], {
29
+
encoding: "utf-8",
30
+
stdio: ["pipe", "pipe", "pipe"],
31
+
}).trim();
32
+
}
33
+
34
+
export async function writeExecutable(
35
+
path: string,
36
+
content: string,
37
+
): Promise<void> {
38
+
await mkdir(dirname(path), { recursive: true });
39
+
await writeFile(path, content, { mode: 0o755 });
40
+
}
41
+
42
+
export async function writeConfig(
43
+
path: string,
44
+
content: string,
45
+
): Promise<void> {
46
+
await mkdir(dirname(path), { recursive: true });
47
+
await writeFile(path, content, { mode: 0o644 });
48
+
}
+14
tsconfig.json
+14
tsconfig.json
···
1
+
{
2
+
"compilerOptions": {
3
+
"target": "es2022",
4
+
"module": "node16",
5
+
"moduleResolution": "node16",
6
+
"outDir": "dist",
7
+
"rootDir": "src",
8
+
"strict": true,
9
+
"esModuleInterop": true,
10
+
"skipLibCheck": true,
11
+
"declaration": true
12
+
},
13
+
"include": ["src"]
14
+
}
+21
verify-templates.ts
+21
verify-templates.ts
···
1
+
import { llamaChatServer, llamaCompleteServer, llamaStart, llamaStop, aiCode, aiAsk, aiPipe } from "./src/templates/scripts.js";
2
+
import { aiderConfig, aiderEnv } from "./src/templates/aider.js";
3
+
4
+
const templates: [string, string][] = [
5
+
["llama-chat-server", llamaChatServer()],
6
+
["llama-complete-server", llamaCompleteServer()],
7
+
["llama-start", llamaStart()],
8
+
["llama-stop", llamaStop()],
9
+
["ai-code", aiCode()],
10
+
["ai-ask", aiAsk()],
11
+
["ai-pipe", aiPipe()],
12
+
["aider.conf.yml", aiderConfig()],
13
+
["aider .env", aiderEnv()],
14
+
];
15
+
16
+
for (const [name, content] of templates) {
17
+
console.log(`\n${"=".repeat(60)}`);
18
+
console.log(` ${name}`);
19
+
console.log("=".repeat(60));
20
+
console.log(content);
21
+
}