https://github.com/silaorg/sila
AI superpowers you own. Sila is an open alternative to ChatGPT where you own AI assistants, chats and data.
https://github.com/silaorg/sila
agents ai chatbot chatgpt claude gemini llms svelte
Last synced: 3 months ago
JSON representation
AI superpowers you own. Sila is an open alternative to ChatGPT where you own AI assistants, chats and data.
- Host: GitHub
- URL: https://github.com/silaorg/sila
- Owner: silaorg
- License: mit
- Created: 2024-04-01T14:23:10.000Z (about 2 years ago)
- Default Branch: main
- Last Pushed: 2026-01-20T08:48:22.000Z (5 months ago)
- Last Synced: 2026-01-20T09:36:03.401Z (5 months ago)
- Topics: agents, ai, chatbot, chatgpt, claude, gemini, llms, svelte
- Language: TypeScript
- Homepage: https://silain.com
- Size: 11.7 MB
- Stars: 13
- Watchers: 1
- Forks: 1
- Open Issues: 12
-
Metadata Files:
- Readme: README.md
- License: LICENSE.md
Awesome Lists containing this project
README
Sila
AI superpowers you own
Download
·
Why did we make it
Sila is like ChatGPT, but you own your assistants, chats, and files. The more you use AI, the more it learns about you and your preferences. This data becomes more valuable over time. That's why you should own it, not OpenAI or Google.
Unlike most similar projects, you don't need to host a server or register an account to use Sila. Even though your data is stored locally as plain files, it can be synced effectively across multiple devices. Later, you can even host your workspace on a server and invite team members. Sila combines the power of local-first with collaborative capabilities.
[Watch how to get started in Sila](https://youtu.be/S-CgZbHXXOw)
## Features
### Workspaces
Organize your chats and assistants into separate workspaces. Each workspace can have its own assistants, files, themes, and languages. Create multiple workspaces for different purposes and switch between them quickly.
### Files and folders
Attach files to chats and reuse them across your workspace. Organize your assets into folders. Sila has its own virtual file system.
### Documents
Assistants can write and edit text documents (in Markdown) locally inside a chat and save them to workspace folders so they can be referenced in other conversations.
### Flexible chats
When you chat, you can switch between assistants, branch conversations, reference files from across the workspace, and edit messages created by you or the AI.
### Assistants
Create your own assistants with their instructions, AI models, and tools. Assistants can support different workflows. For example, an assistant can always search a specific source, reference a document, and create conversations in a chosen project.
### Local-first
Workspaces are stored locally. You can sync them with iCloud, Dropbox, or similar services to use across devices. Your assistants, chats, and generated data remain under your control. Sync conflicts are resolved automatically. No accounts required. Sila works offline if the AI model runs on your device.
### Tabs like in VSCode
Switch between tabs and split windows. It works much like VS Code. You can have multiple conversations open across tabs, chat with different assistants at the same time, and quickly switch between them.
### Many themes
Use different themes for your workspaces—from colorful to minimal. It's a simple way to set a mood or tell your workspaces apart.
### Any AI models
You can use any major AI model from the ones powering ChatGPT, to Claude and Gemini.
### No subscriptions
Pay as you go for provider API costs or local compute if you run models yourself.
## Download the app
**Download Sila:**
- [Download for macOS](https://silain.com/download#platforms)
- [Download for Windows](https://silain.com/download#platforms)
- [Download for Linux](https://silain.com/download#platforms)
## Documentation
### Product
Learn more about Sila's features in the [product documentation](https://silain.com/v1/features/).
### Development
Want to build from source? Check out the [quick start guide](https://silain.com/v1/dev/quick-start) and [development documentation](https://silain.com/v1/dev/) covering the architecture, codebase, testing, and contribution guidelines.
## Tech stack
- Standalone application (desktop + mobile)
- Desktop app runs on [Electron](https://www.electronjs.org/)
- Mobile app runs on [Capacitor](https://capacitorjs.com/)
- [TypeScript](https://www.typescriptlang.org/) as the main language
- Frontend uses [Svelte 5 + SvelteKit](https://svelte.dev/)
- Everything runs locally (no server yet) plus external APIs
- Styling via [Tailwind CSS](https://tailwindcss.com/)
- Components from [Skeleton design system](https://github.com/skeletonlabs/skeleton)
- Inference with AI is done through [AIWrapper](https://github.com/mitkury/aiwrapper)
- Sync handled by [RepTree](https://github.com/mitkury/reptree)
- Tiling tabs like in VSCode are built with [TTabs](https://github.com/mitkury/ttabs)
- Context for AI agent generated with [Airul](https://github.com/mitkury/airul)
## Related projects
Sila is built alongside several companion projects maintained by Sila's authors:
- **AI inference** - [AIWrapper](https://github.com/mitkury/aiwrapper)
- **Info about AI models** - [AIModels](https://github.com/mitkury/aimodels)
- **Tiling tabs** - [TTabs](https://github.com/mitkury/ttabs)
- **Sync** - [RepTree](https://github.com/mitkury/reptree)
- **AI context** - [Airul](https://github.com/mitkury/airul)