Rytriv puts an AI agent in VS Code that reads your codebase before it answers. Rather than guessing from whatever happens to fit in a prompt, it runs retrieval over your project, pulls the specific code that matters, and works from that.
How it works
Rytriv indexes your workspace on your machine and keeps that index live while you type. When you ask something or hand off a task, it gathers the relevant files, symbols, and call paths and gives the agent a focused, accurate slice of your project to reason over. Answers point back to the code they came from, so you can check the source instead of trusting a guess.
What you can do with it
- Ask and plan in chat. Questions like "where does auth happen", "what breaks if I rename this", or "sketch a migration plan" come back with references to the real files behind the answer.
- Hand work to the agent. It reads, edits across many files, runs the terminal, searches the web, drives a browser, and runs your tests, then reviews its own output before handing it back.
- Keep long sessions honest. Rytriv records what gets decided as a conversation grows, and warns you when a later answer drifts from something you already settled.
- Bring your own tools. Connect them over MCP and the agent can call them mid task.
Setup takes a minute
- Install Rytriv.
- Open the Rytriv panel and click Sign in. Approve in the browser and you land back in the editor.
- Open a project. Indexing starts on its own.
A model comes with your plan, so there is nothing to paste or configure. Prefer to run on your own account? Switch on Bring Your Own Key and add your key.
More at rytriv.com.