v2.1.0

Dashboard simplification, Tavily, Exa, YouTube, MCP, and runtime improvements.

by @saifyxpro

Released on March 16, 2026

Dashboard and Playground


  • Removed the old profile-based session model and simplified runtime behavior around one shared global browser configuration.
  • Streamlined the dashboard shell, sidebar, settings surfaces, and playground headers to use a more consistent layout and flatter interaction model.
  • Added the Tavily, Exa, and YouTube workspaces to the playground and aligned them with the newer shared workbench shell.
  • Added loading skeletons and refined sidebar hover, collapse, and system load presentation across the dashboard.

Website, Search, and Jobs


  • Expanded the website workflows with better discovery, queued jobs, improved progress handling, and cleaner website naming across the UI and API.
  • Added queue-backed jobs and simplified proxy settings so long-running workflows have a clearer status model.
  • Added Tavily search and research support, then followed up with interaction fixes and workbench polish.
  • Added Exa search support and unified the shared search workspace shell.

YouTube and MCP


  • Added the new YouTube engine integration with unified extract and save flows inside the playground.
  • Simplified the supported YouTube client families and improved the shared loading and results behavior around that flow.
  • Added a remote MCP endpoint to the API with secured API-key-based access and job scoping.

Deployment and Runtime


  • Improved Docker startup by booting with Prisma migrations automatically and by avoiding misleading missing-env warnings when environment values are already injected.
  • Routed dashboard proxy behavior through an internal API URL for container networking.
  • Hardened runtime service configuration so queue, markdown, and engine endpoints follow environment-driven configuration more consistently.
  • Standardized the root development workflow around shared pnpm, Nx, and mise entrypoints.

Documentation


  • Added setup and API guides, refreshed README scraper visuals, and aligned scraper run actions with the current product surface.