Oloo Agent

Changelog

Release notes for Oloo Agent. We ship when it is ready, not on a fixed schedule.

v0.1.0-preview

Technical Preview

March 2026 — First public release

What works today

Observe

  • Page content extraction and entity recognition on most sites
  • Full-page screenshots and DOM snapshot capture

Learn

  • Multi-model AI chat with streaming responses (DeepSeek, OpenAI, Claude, Gemini, and 4 more providers)
  • Structured content summarization and analysis

Organize

  • Conversation history stored locally with full persistence
  • Skill system: create, import, and run reusable AI workflows from Markdown files
  • Trigger system: auto-run skills on pages matching your URL patterns
  • Hook system: intercept lifecycle events for custom behavior

Outgrow

  • 41 browser automation tools via Chrome DevTools Protocol (click, type, fill, screenshot, extract text, execute JS, navigate, scroll, and more)
  • Settings and API key management for all 8 providers
  • Chrome MV3 side panel interface

In progress

  • Knowledge base and knowledge graph (Organize phase) — planned for v0.2.0
  • Agent task decomposition (multi-step autonomous Outgrow) — planned for v0.2.0
  • Sub-agent orchestration (parallel task execution) — planned for v0.3.0
  • Chrome Web Store listing — in preparation
  • Full internationalization framework — English UI available, structured i18n coming
  • Dark mode — planned for v0.2.0

Not started yet

  • Vector search and semantic retrieval for knowledge base
  • Full-page translation
  • Team and shared skill libraries
  • Mobile companion (not planned — Chrome desktop only)

More versions will appear here as the project progresses. Follow development on GitHub.