About Codecora
Codecora builds developer tools for the AI agent era — open-source at heart, with room to grow.
Our Philosophy
Local-first by default. Your data should stay on your machine. We build tools that work offline, need zero API keys, and never phone home.
Simple to start, powerful to scale. Single binary installs, sensible defaults, and configurable enough for any workflow — from a solo dev to a team of AI agents.
Open-source core. Our tools are MIT or Apache 2.0 licensed. We believe in transparent, community-driven development — and we may offer paid tiers for teams and enterprise in the future.
The Stack
All of our tools are built with Rust for performance and reliability:
| Project | Role | Why Rust? |
|---|---|---|
| Cora CLI | Code review | Native speed, zero cold starts, cross-platform |
| Cora Review Action | CI/CD integration | GitHub Action wrapper for Cora CLI |
| Uteke | Semantic memory | 30ms recall, single binary, ONNX runtime |
| TrapFall | Error capture | Minimal footprint, 6MB Docker, MUSL static |
| Rungu | Feedback board | Lightweight self-hosted, zero dependencies |
The Problem We Solve
AI agents are getting powerful — but they're forgetful, blind, and noisy:
- Code review is slow and expensive → Cora makes it instant and configurable
- Agents forget context across conversations → Uteke gives them persistent memory
- Error monitoring is bloated → TrapFall fits in 6MB with full Sentry compatibility
- User feedback is scattered → Rungu keeps it in one self-hosted place
Contact & Community
- GitHub: github.com/codecoradev
License
Each project has its own license:
- Cora CLI: MIT
- Cora Review Action: MIT
- Uteke: Apache 2.0
- TrapFall: Apache 2.0
- Rungu: MIT