Project Templates์ถ์ฒ: Show HN์กฐํ์ 11
Show HN: Claude-relais โ A plan/build/judge loop mixing Claude with Cursor
By crog2026๋
2์ 16์ผ
**Show HN: Claude-relais โ A plan/build/judge loop mixing Claude with Cursor**
I built claude-relais to stop throwing expensive reasoning models at every line of code (I can't spend 500$/month on AI subs).The idea: Claude acts as orchestrator: it plans tasks, breaks them into bounded units, and judges the output. Cursor agents (basic plan, almost unlimited and super fast) handle the actual code generation, which doesn't need frontier-level reasoning.A guarded loop (PLAN -> BUILD -> JUDGE) enforces safety constraints between steps (no destructive ops, scoped file access, bounded iterations).In practice, this brings my AI-assisted coding cost to ~$40/month while keeping the quality of having a strong model in the loop for architecture and review decisions.The core tradeoff: you give up the simplicity of one model doing everything, but you gain cost control and a natural separation between "thinking" and "typing."Repo: https://github.com/clementrog/claude-relaisWould love feedback on the orchestration approach, especially if others have tried similar multi-model setups and hit failure modes I haven't seen yet. Comments URL: https://news.ycombinator.com/item?id=47026886 Points: 1 # Comments: 0
---
**[devsupporter ํด์ค]**
์ด ๊ธฐ์ฌ๋ Show HN์์ ์ ๊ณตํ๋ ์ต์ ๊ฐ๋ฐ ๋ํฅ์ ๋๋ค. ๊ด๋ จ ๋๊ตฌ๋ ๊ธฐ์ ์ ๋ํด ๋ ์์๋ณด์๋ ค๋ฉด ์๋ณธ ๋งํฌ๋ฅผ ์ฐธ๊ณ ํ์ธ์.
I built claude-relais to stop throwing expensive reasoning models at every line of code (I can't spend 500$/month on AI subs).The idea: Claude acts as orchestrator: it plans tasks, breaks them into bounded units, and judges the output. Cursor agents (basic plan, almost unlimited and super fast) handle the actual code generation, which doesn't need frontier-level reasoning.A guarded loop (PLAN -> BUILD -> JUDGE) enforces safety constraints between steps (no destructive ops, scoped file access, bounded iterations).In practice, this brings my AI-assisted coding cost to ~$40/month while keeping the quality of having a strong model in the loop for architecture and review decisions.The core tradeoff: you give up the simplicity of one model doing everything, but you gain cost control and a natural separation between "thinking" and "typing."Repo: https://github.com/clementrog/claude-relaisWould love feedback on the orchestration approach, especially if others have tried similar multi-model setups and hit failure modes I haven't seen yet. Comments URL: https://news.ycombinator.com/item?id=47026886 Points: 1 # Comments: 0
---
**[devsupporter ํด์ค]**
์ด ๊ธฐ์ฌ๋ Show HN์์ ์ ๊ณตํ๋ ์ต์ ๊ฐ๋ฐ ๋ํฅ์ ๋๋ค. ๊ด๋ จ ๋๊ตฌ๋ ๊ธฐ์ ์ ๋ํด ๋ ์์๋ณด์๋ ค๋ฉด ์๋ณธ ๋งํฌ๋ฅผ ์ฐธ๊ณ ํ์ธ์.