Overview
Jeremy Daer from 37signals explains why AI agents should use CLI tools instead of REST APIs. CLI interfaces enable cheaper, faster models to handle complex tasks that would otherwise require expensive, powerful models for API integration.
Key Facts
- CLI tools improve accuracy and success rate for complex operations - agents can reliably handle pagination, rate limits, and auth failures
- Cheaper, fast models (gpt-5-nano, haiku-4.5) can use CLI tools effectively - dramatically reduces operational costs for autonomous agents
- Raw APIs require expensive ‘strong’ models and multiple iterations - autonomous agents burn through costly tokens doing repetitive work
- CLI approach saves context window space - agents can operate more efficiently within token limits
Why It Matters
This matters because it reveals a practical strategy for making AI agents economically viable - using simpler interfaces that allow cheaper models to perform complex tasks reliably, rather than burning expensive tokens on API complexity.