April 10, 2026
Bots code, devs clap back
Launch HN: Twill.ai (YC S25) – Delegate to cloud agents, get back PRs
Bots write your PRs, devs ask: “cool or cash grab”
TLDR: Twill.ai launched cloud coding agents that open tested, ready-to-review code changes and plug into GitHub, Slack, and more. The thread split between believers in 24/7 bot coworkers and critics demanding benchmarks and calling it a pricey middleman versus just using Claude directly.
Startup Twill just rolled into Launch HN promising a tireless teammate: cloud coding agents that study your repo, write fixes and small features, run tests in their own sandboxes, and open PRs (code change requests) ready for review. You can pick from familiar bots (Claude Code, OpenCode, Codex), even run several at once, and it plugs into GitHub, Linear, and Slack. There’s a Free tier (10 runs), Pro at $50, and Max at $200, plus free for open source. The pitch: fewer chores, faster shipping, no local setup—just mention @twill and let the robots cook.
The comments lit up. One camp is chanting future now, with a user declaring 24/7 coding agents are inevitable, especially for companies that want on‑prem or private cloud. Another camp went straight for the wallet: a top jab asked if this is just a middleman charging to run Claude on your behalf. Skeptics also want receipts—benchmarks for the claim that rerunning the same agent boosts success. The counter‑vibe: builders who’ve shipped “well over 1k PRs” with an open‑source cousin swear the prompt‑to‑PR workflow is liberating and love extras like session videos. Meme patrol: “intern that never sleeps,” “bots on call, humans on vacation,” and the eternal “will it actually merge?” The mood: hype vs. markup vs. measurable results.
Key Points
- •Twill.ai offers cloud-based coding agents that plan, implement, test, and open verified PRs to automate routine engineering tasks.
- •Users can choose from agents like Claude Code, OpenCode, and Codex, run multiple agents in parallel, or rerun an agent to improve success rates.
- •Agents build and test changes in isolated sandbox environments with secure access to logs and ports; developers can SSH in using IDEs like Cursor or VS Code.
- •The platform integrates with GitHub, Linear, and Slack for task initiation via @twill mentions, minimizing context switching.
- •Pricing includes Free (10 runs), Pro ($50 for 50 runs), and Max ($200 for 250 runs, larger machines), with open source free; enterprise offers BYO API keys and private VPC deployment.