Welcome to Quira
Quira is a local-first browser that transforms your browsing activity into a structured, queryable knowledge graph. Unlike traditional browsers that treat tabs as disposable, Quira automatically captures the context of your research — what you read, how topics connect, and where your investigation left off — so you can ask your browser anything and get answers grounded in your own experience.
This documentation covers everything you need to get started with Quira, understand its core concepts, build on its APIs, and contribute to its development.
New to Quira?
Start with the Quick Start guide to get up and running in under 5 minutes, then explore Core Concepts to understand how Quira structures your browsing context.
Explore the documentation
Install Quira, configure your local AI model, and start building your Context Graph in minutes.
Understand the Context Graph, Spaces, local AI integration, and how Quira maps your research flow.
Set up the development environment, learn the code conventions, and submit your first pull request.
Explore the Plugin API, WebExtension compatibility layer, and the Graph Query API for building extensions.
What's new
Context Graph + Research Replay
Full Context Graph with WebGPU rendering, entity extraction, and the ability to resume past research sessions exactly where you left off.
Local AI + Natural Language Query
On-device Phi-3.5 Mini integration with Command Palette NL Query. Ask your browser anything and get answers grounded in your own browsing history.
Gecko Fork + Context Spaces
Initial Gecko-based browser with full WebExtension compatibility (MV2 + MV3), Context Spaces for tab organization, and Chrome-compatible UI presets.
Architecture at a glance
+------------------------------------------------+
| Quira Browser |
+-------------+-------------------+---------------+
| UI Layer | Context Layer | Agent Layer |
| (Rust/GPU) | (Local Graph DB) | (Local LLM) |
+-------------+-------------------+---------------+
| Gecko-based Rendering Engine |
| (Phase 1: Gecko -> Phase 3: Servo) |
+------------------------------------------------+
| WebGPU | WASM | Post-Quantum TLS |
+------------------------------------------------+| Layer | Technology | Rationale |
|---|---|---|
| Rendering | Gecko (Phase 1) | Full web compatibility, WebExtensions support, ship-ready |
| UI | Rust + WebGPU | Native performance, low memory footprint |
| Context Graph | SQLite + FTS5 + sqlite-vec | Zero dependencies, offline-first, single-file database |
| Local LLM | llama.cpp / MLX | GPU acceleration, privacy-preserving |
| Extensions | WebExtensions (MV2+MV3) | Broad ecosystem compatibility, including MV2 extensions |
| Sync | CRDT-based P2P | Serverless, conflict-free, no cloud dependency |
Join the community
Quira is open source and community-driven — here's how to get involved:
- GitHub Repository — Star the project, file issues, and submit pull requests
- Discord Server — Chat with the team and other contributors
- RFC Process — Propose and discuss significant changes
- Roadmap — See what's planned and what's in progress