Blog
Product updates, and lessons on building AI agents that work like team members.
All posts
- Your agents don't belong in your codebase Learn how we've decoupled "how agents run" from "what agents are" (their system prompt, skills, tools) to improve our product iteration velocity
- videostil: Bringing Video Understanding to Every LLM Turn videos into structured LLM-ready data with configurable frame extraction, smart deduplication, and visual debugging — all open source.
- Appwright: A new test framework for mobile apps Appwright is a single package that integrates a mobile automation driver and a test runner for end-to-end testing of mobile apps.