Recall 2.0 brings agentic knowledge chat

Recall is moving from “AI bookmark manager” toward a personal RAG layer that can follow users across apps, which is the right direction if second-brain tools want to survive ChatGPT’s built-in memory push.

• –The strongest feature is mixed context: saved knowledge plus internet search gives answers grounded in both personal history and current information.
• –MCP and API access make this more interesting for developers, because Recall can become a private retrieval source for agents instead of just another standalone notes app.
• –Model switching is a practical hedge against fast-changing LLM quality and pricing, especially for users who care about cost, reasoning depth, or provider preference.
• –The open question is retrieval quality: personal knowledge chat only feels magical if it cites the right saved items and avoids confidently blending stale notes with fresh web results.