Context Zero Engine
Local code-intelligence engine for AI agents — indexes a repository into a code graph and serves context over MCP and HTTP.
We build for people, to make what they use secure and reliable.
Every product on its own page — what it does, where it runs, how to get it. Install, open, it works.
Say it or type it: REX writes and runs code, works the terminal, drives the browser and carries long jobs to the end. In private testing — no public build yet.
The waitlist hears first.
Open problems and in-progress builds, documented before they become products.
Looking at new platforms to build for — no dates yet.
Research into platforms we might build for. Nothing is real until there’s something you can download, and no launch date is implied.
Read the full noteFree apps and open-source tools — no account, no ads, no catch.
Local code-intelligence engine for AI agents — indexes a repository into a code graph and serves context over MCP and HTTP.
CPU-only Rust LLM inference engine — 4-bit quantized kernels and local serving, no CUDA or PyTorch.
Self-hosted credential firewall for AI agents — your agent sees placeholder keys while a deny-by-default proxy injects the real ones.
Free, open-source live speech translation in your ear on Android — bring your own Gemini or OpenAI key.
Free Android document reader with PDF search and page-text copy.
What has to be true before anything here gets called a product.
It gets called a product when it holds up under real use. Not before.
Work in progress, labeled as work in progress.
Source or download available. That is what public means — not that it is finished.
It works and it still does not ship. Working is the floor.