Launch HN: Twill.ai (YC S25) – Delegate to cloud agents, get back PRs
6 hours ago
- #AI coding assistant
- #automated development workflow
- #sandbox testing
- Twill is an AI agent that autonomously writes code, runs tests, fixes failures, and opens PRs, only requiring human input when needed.
- It follows a structured workflow with fixed steps: research, planning, human approval, implementation in a sandbox, AI code review, and merging.
- Benefits include increased productivity with automated bug fixes, dependency updates, and docs, enabling small teams to ship like larger companies.
- Users can assign tasks to various coding agents like Claude Code, OpenCode, and Codex, run them in parallel, and compare outputs.
- Agents verify changes in isolated dev environments, spin up sandboxes for testing, and integrate with tools like GitHub, Linear, and Slack.
- Twill allows starting from proven automation templates to run recurring engineering workflows seamlessly.