v1.6.1 Release

New features

• Richer page understanding: detects interactive elements and complex DOMs (iframes, shadow DOM)
• Choose scraping output: markdown, JSON, or images
• Added model providers: Together AI and Anthropic

Improvements

• Pick your perception mode per session: simpler observation output
• Easier persona/vault setup
  • Flexible vault/persona APIs; pass actions as dicts
• Health checks and clearer error messages

▸ Explore the releases here