LangChain launches OpenWiki Brains: proactive memory framework for agents across email, Notion, GitHub, web

LangChain released OpenWiki Brains, a local-first memory framework that allows AI agents to build proactive context from connected data sources. Unlike reactive memory systems (built-in agent recall of what users explicitly share), OpenWiki Brains continuously fetches and synthesizes information from sources like Gmail, Notion, GitHub, Twitter/X, Hacker News, and web search, then auto-refreshes a local wiki that agents can reference. Personal Brain mode is the main launch feature, creating a structured knowledge base around active projects, people, saved links, and recent notes.

The framework introduces "connectors" that pull data from external sources. Deterministic connectors (Gmail, Twitter/X, GitHub) fetch recent items directly; agentic connectors (Notion, web search) use LLM-driven search with user-specified focus areas. The local-only design means wikis can refresh on a user-defined schedule without requiring server infrastructure, running scheduled jobs on the machine directly. LangChain also supports Code Brain mode, the original use case for generating and maintaining documentation inside git repos.

This shifts agent design from "I'll call tools each time you ask" to "I maintain a live context layer." OpenWiki's connectors cover workplace communication (Gmail, Slack coming soon), note-taking (Notion), code (GitHub), and web discovery (Twitter/X, HN), addressing the scattered nature of context across tools. Teams can now spin up agents with fresh context without manual prompt engineering or copy-paste workflows.

For developers building agents: OpenWiki Brains lower the operational burden of keeping agent context current—a critical blocker when agents need to act on timely information (project updates, recent research, latest code changes). The local-first model also appeals to privacy-conscious teams and those avoiding vendor lock-in on memory infrastructure.