A local-first private AI assistant for everyday use. Runs on-device models with encrypted P2P sync, and supports sharing chats publicly on ATProto.
10
fork

Configure Feed

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

Update README

Updated logo display for dark and light modes, refined installation instructions, and clarified design partner outreach.

authored by

Ankesh Bharti and committed by
GitHub
39f8731c a6c94031

+13 -10
+13 -10
README.md
··· 1 1 <!-- LOGO --> 2 2 <p align="center"> 3 3 <a href="https://github.com/tileshq/"> 4 - <img 5 - src="https://avatars.githubusercontent.com/u/210493283?s=400&u=1ee6e44b6a683b16bdb6e9e853c7ebd8c7fd4268&v=4" 6 - alt="Tiles Logo" 7 - width="128" 8 - /> 4 + <picture> 5 + <source media="(prefers-color-scheme: dark)" srcset="https://avatars.githubusercontent.com/u/210493283?s=400&u=1ee6e44b6a683b16bdb6e9e853c7ebd8c7fd4268&v=4"> 6 + <source media="(prefers-color-scheme: light)" srcset="https://github.com/user-attachments/assets/677d3d63-a9b0-4e92-8c7d-f39ceffc1b0d"> 7 + <img src="https://github.com/user-attachments/assets/677d3d63-a9b0-4e92-8c7d-f39ceffc1b0d" alt="Tiles Logo" width="128"> 8 + </picture> 9 9 </a> 10 10 </p> 11 11 ··· 30 30 31 31 The CLI is the fastest way to get started and will feel familiar if you have used tools like Ollama or LM Studio. 32 32 33 - Install and run: 33 + Install the signed macOS package: 34 + 35 + https://download.tiles.run/tiles-0.4.3-signed.pkg 36 + 37 + Then run: 34 38 35 39 ```bash 36 - curl -fsSL https://tiles.run/install.sh | sh 37 40 tiles 38 41 ``` 39 42 ··· 42 45 Tilekit is a Rust-based SDK for customizing local models and agent experiences within Tiles. 43 46 44 47 It provides: 45 - - A Modelfile specification for defining and sharing models 46 - - Fast, efficient local deployment across consumer platforms 48 + - Modelfile specification and tooling for working with models 49 + - Fast, efficient local deployment of models across consumer platforms 47 50 - Built on open source specifications such as Open Responses API 48 51 - Model composition and chaining, with MIR support currently in development 49 52 ··· 62 65 63 66 Tiles is built for privacy conscious users who want intelligence without renting their memory to centralized providers. 64 67 65 - We are seeking design partners for training workloads that align with our goal of a verifiable privacy perimeter. Contact us at hello@tiles.run. 68 + We are seeking design partners for TEE based cloud workloads that align with our goal of a verifiable privacy perimeter. Contact us at hello@tiles.run. 66 69 67 70 ## Contributing 68 71