personal memory agent
0
fork

Configure Feed

Select the types of activity you want to include in your feed.

docs: README + INSTALL trinity rewrite per system-anatomy canon

Replace surveillance vocabulary in owner-facing prose with the
canon's trinity framing — solstone = observers + sol agent + journal
in canonical order. Code-adjacent uses of `capture` (ASCII diagram,
CLI verb comments, engineering doc-pointer table row) preserved.

Precedent: ddec60d5 (AGENTS.md §13 system-anatomy canon pointer,
lode kewybdze). Canon source: ~/projects/extro/cmo/brand/system-anatomy.md.

README.md edits (11 lines):
- L5 headline: capture frame → trinity (observers/sol/journal).
- L7 paragraph: "recording audio and screen activity" → "experiencing
your day along with you" — paragraph now traces trinity across its
three sentences.
- L13 caption: "passive capture" → "observations".
- L19 bullet: "continuous audio capture (via standalone observers)"
→ "standalone observers take in audio continuously".
- L20 / L23 bullets: "entity tracking" / "commitment tracking" →
noun-phrase labels "people and projects" / "commitments"; "tracked"
→ "remembered"; "captured from natural conversation" → "extracted
from natural conversation".
- L32 caption: "screen capture analysis" → "screen analysis".
- L34 alt text + L36 caption: "entity tracking" → "people and projects";
"tracked" → "remembered".
- L62 architecture bullet: "receives captured audio and screen activity"
→ "receives audio and screen observations"; "WebM screen recordings"
→ "WebM screen media"; pre-existing typo "vian" → "via" fixed in
passing.
- L63 architecture bullet: "screen captures" → "screen observations".

INSTALL.md edits (3 lines):
- L3 opening: capture frame → trinity ("solstone is your co-brain:
observers experience your day along with you, sol curates your
memories, and your journal holds everything").
- L99: "doesn't capture anything on its own" → "doesn't observe
anything on its own".
- L133: "captures screen and audio continuously" → "observes screen
and audio continuously".

Preserved by design (canon permits `capture` in code-adjacent contexts):
- README L17 "system of record" — deliberate noun-phrase rhetorical
contrast with "system of intelligence", not a surveillance verb.
- README L43 ASCII diagram cell `| capture |` — internal architecture
vocabulary.
- README L94 code comment `# Start the full stack (capture + processing
+ web)` — CLI / pipeline label.
- README L98 code comment `# Capture a screenshot of the web UI` — CLI
verb (sol screenshot).
- README L109 documentation table row `| Capture pipeline |
docs/OBSERVE.md |` — engineering-doc pointer.

Out of scope (deferred to follow-on lodes):
- think/sol_cli.py strings, apps/*/templates, convey/templates,
settings copy.
- docs/INSTALL.md and other docs/*.md (engineering references where
capture is canon-permitted; INSTALL.md is the owner-facing entry).
- README "your co-brain" headline phrasing unification with
solstone.app landing page (Phase D, CMO-CTO follow-on).

CI note: `make ci` blocks on a pre-existing `onnxruntime`
ModuleNotFoundError in the lode venv (observe/transcribe/overlap.py
import in the .installed bootstrap), unrelated to this markdown diff.
The relevant gate scripts pass when invoked directly:
- scripts/gate_agents_rename.py → exit 0
- scripts/check_layer_hygiene.py → exit 0

+14 -14
+3 -3
INSTALL.md
··· 1 1 # installing solstone 2 2 3 - these instructions are for a coding agent and human working together. solstone is an AI life journal — it captures everything you see and hear on your computer, processes it with AI, and gives you superhuman memory. open source, made by sol pbc. 3 + these instructions are for a coding agent and human working together. solstone is your co-brain: observers experience your day along with you, sol curates your memories, and your journal holds everything. open source, made by sol pbc. 4 4 5 5 **supported platforms:** linux (primary), macOS. windows is not yet supported. 6 6 ··· 96 96 97 97 ## install an observer 98 98 99 - solstone doesn't capture anything on its own — it needs an observer for the platform. 99 + solstone doesn't observe anything on its own — it needs an observer for the platform. 100 100 101 101 ```bash 102 102 uname -s ··· 130 130 131 131 ## done 132 132 133 - once the observer is running, solstone captures screen and audio continuously, transcribes conversations, extracts people and projects, builds a knowledge graph, and makes everything searchable at http://localhost:5015. all data stays in `journal/` directories — one per day. 133 + once the observer is running, solstone observes screen and audio continuously, transcribes conversations, extracts people and projects, builds a knowledge graph, and makes everything searchable at http://localhost:5015. all data stays in `journal/` directories — one per day. 134 134 135 135 source code: https://github.com/solpbc/solstone 136 136 company: https://solpbc.org
+11 -11
README.md
··· 2 2 3 3 # solstone 4 4 5 - your co-brain — captures everything you see and hear, processes it with AI, and gives you superhuman memory. 5 + your co-brain — observers experience your day along with you, sol curates your memories, and your journal holds everything. 6 6 7 - solstone runs in the background on your computer, recording audio and screen activity. AI agents transcribe, extract entities, detect meetings, build knowledge graphs, and surface daily insights — all without any manual input. everything stays on your machine in daily journal directories. open source, local-first, no cloud required. 7 + solstone runs in the background on your computer, experiencing your day along with you. AI agents transcribe, extract entities, detect meetings, build knowledge graphs, and surface daily insights — all without any manual input. everything stays on your machine in daily journal directories. open source, local-first, no cloud required. 8 8 9 9 Python 3.10+, Linux + macOS, AGPL-3.0-only, maintained by [sol pbc](https://solpbc.org). 10 10 11 11 <img src="docs/static/screenshot-home.png" alt="solstone daily dashboard" width="800"> 12 12 13 - *Daily dashboard — goal, todos, upcoming events, and detected entities, all generated from passive capture. Facet tabs organize your life by project or context.* 13 + *Daily dashboard — goal, todos, upcoming events, and detected entities, all generated from observations. Facet tabs organize your life by project or context.* 14 14 15 15 ## what you get 16 16 17 17 **a system of intelligence, not just a system of record.** 18 18 19 - - **automatic transcription** — continuous audio capture (via standalone observers) with speaker identification. every conversation, transcribed and searchable. 20 - - **entity tracking** — people, companies, and projects extracted from your conversations and tracked across time. 19 + - **automatic transcription** — standalone observers take in audio continuously with speaker identification. every conversation, transcribed and searchable. 20 + - **people and projects** — extracted from your conversations and remembered across time. 21 21 - **knowledge graphs** — relationships between entities mapped automatically. who works with whom, which projects connect to which people. 22 22 - **meeting detection** — meetings identified, summarized, and linked. meeting prep that surfaces what you discussed last time and personal context you'd forget. 23 - - **commitment tracking** — todos captured from natural conversation. no manual entry. 23 + - **commitments** — todos extracted from natural conversation. no manual entry. 24 24 - **facet organization** — group everything by project or context (work, personal, client-name) with scoped views across all apps. 25 25 - **AI chat** — talk to your journal. ask anything about your digital life and get answers grounded in your actual data. 26 26 - **full-text search** — find anything you've ever seen or heard. ··· 29 29 30 30 <img src="docs/static/screenshot-transcripts.png" alt="solstone transcript viewer" width="800"> 31 31 32 - *Transcript viewer — dual-timeline navigation, speaker-diarized dialogue, audio playback, screen capture analysis. every conversation browsable by time.* 32 + *Transcript viewer — dual-timeline navigation, speaker-diarized dialogue, audio playback, screen analysis. every conversation browsable by time.* 33 33 34 - <img src="docs/static/screenshot-entities.png" alt="solstone entity tracking" width="800"> 34 + <img src="docs/static/screenshot-entities.png" alt="solstone people and projects" width="800"> 35 35 36 - *Entity tracking — people, companies, and projects automatically extracted and tracked across your journal with mention counts and relationship data.* 36 + *People and projects — automatically extracted and remembered across your journal with mention counts and relationship data.* 37 37 38 38 ## architecture 39 39 ··· 59 59 +-------------+ 60 60 ``` 61 61 62 - - **observe** — receives captured audio and screen activity from standalone observers (solstone-linux, solstone-tmux, solstone-macos) vian observer ingest. processes FLAC audio, WebM screen recordings, and timestamped metadata. 63 - - **think** — transcribes audio (faster-whisper), analyzes screen captures, extracts entities, detects meetings, and indexes everything into SQLite. runs 30 configurable agent/generator templates from `talent/`. 62 + - **observe** — receives audio and screen observations from standalone observers (solstone-linux, solstone-tmux, solstone-macos) via observer ingest. processes FLAC audio, WebM screen media, and timestamped metadata. 63 + - **think** — transcribes audio (faster-whisper), analyzes screen observations, extracts entities, detects meetings, and indexes everything into SQLite. runs 30 configurable agent/generator templates from `talent/`. 64 64 - **cortex** — orchestrates agent execution. receives events, dispatches agents, writes results back to the journal. 65 65 - **callosum** — async message bus connecting all services. enables event-driven coordination between observe, think, cortex, and convey. 66 66 - **convey** — Flask-based web interface with 17 pluggable apps for navigating journal data.