Run coding agents in your own workspace. QQ connects to your existing Codex subscription or supported model APIs, with a terminal UI, file editing, shell commands, MCP tools and visible activity. Download one open-source binary and start with a repository. We are adding JEV checkpoints to review tool results and task completion; that integration is in testing and is not yet included in v0.1.0.
What became possible in your product with Astra that was not practical before?
Maker
Astra helped our small team build and test the agent and its checking workflow together: repository tasks, a JEV companion for evidence reviews, and regression tests for failures and cancellation.
QQ lets developers work in their own repository with Codex sign-in. JEV adds a separate assessment of supplied evidence, rather than treating the coding agent's final answer as proof.
The practical change was being able to iterate across implementation, real runs and failure tests in one workflow. We could investigate what the agent actually did and turn failures into test cases while developing the product.
Report
Maker
📌
QQ lets you run coding agents in your own project using your existing Codex subscription. Give it a task, follow its progress, and review the changes.
We’re adding JEV checks throughout the workflow to catch incomplete work before the agent calls it done.
QQ lets you run coding agents in your own project using your existing Codex subscription. Give it a task, follow its progress, and review the changes.
We’re adding JEV checks throughout the workflow to catch incomplete work before the agent calls it done.
What would you try first?