Distillation Labs builds developer tools for the practical efficiency layer around AI. Contextro compresses repository context before an agent edits. Agentyc gives the agent a deterministic browser runtime when it needs the live application.
Shared thesis
Both products are designed around the same constraint: useful AI gets more practical when the system retrieves less junk and returns a tighter, more inspectable surface.
Products
Repository intelligence
Single Rust binary. Local MCP server for Claude, Cursor, Windsurf, and other clients - search code by meaning, trace call graphs, inspect change impact, query git history, and keep memory across sessions without sending code to the cloud.
Browser automation
Deterministic, MCP-first browser automation for coding agents. Launch or attach to Chrome over CDP, inspect compact page state with stable refs, extract structured content, and debug real browser traffic without an LLM in the critical path.
Shared thesis
Both products are designed around the same constraint: useful AI gets more practical when the system retrieves less junk and returns a tighter, more inspectable surface.
Contextro solves the repository side of that problem with compact code retrieval, graph-aware inspection, and durable project memory. Agentyc solves the browser side with stable refs, compact page state, deterministic extraction, and CDP-native debugging.
Where each product fits
Reach for Contextro when the agent needs the right symbol, call path, architecture boundary, or historical context before it edits a repository.
Reach for Agentyc when the agent needs to inspect a rendered UI, act in the browser, wait on network activity, or debug the app from the browser side.
Together they let an agent move from repo understanding to live-browser validation without switching to noisy, full-context tooling.
Next step
Start with the product page when you want the operating model and tradeoffs. Move to the docs when you are ready to install, configure an MCP client, and use the tools in real workflows.