Capy

Capy is an AI software engineering workspace where you can hand it engineering tasks and have AI agents plan, implement, review, and produce pull requests. It’s aimed at teams that want faster throughput on day-to-day work like bug fixes and migrations, without building custom agent workflows from scratch.

• Parallel coding agents for planning, building, and shipping from a single task
• Bug triage and fixes that generate PRs directly from the issues you’re tracking
• Inbuilt review agent that runs as part of PR workflow to reduce review bottlenecks
• E2E testing runs using native app and browser use, with annotated videos
• Automations that trigger work on events or cron, including refactors, docs, and security audits
• Web UI session management to organize and manage multiple agent runs
• Model-agnostic runs (you can run different providers/models for a task)
• Integrations such as Slack, Linear, and Sentry-style inputs, plus an API to start threads and request reviews

What stands out is the full “task to merged PR” framing. You can ask for work like clearing batches of tickets, handling migrations (example shown moving a Next.js SPA to TanStack Start), or running scheduled security audits, and Capy is set up to keep the flow inside your normal code review surface (PRs) rather than just producing a patch somewhere.