activeloopai/hivemind
Hivemind turns your traces into reusable skills across agents
What it solves
Hivemind provides a shared, cloud-backed memory system for AI coding agents. It prevents team members from repeating the same mistakes or re-solving the same problems by allowing agents to share knowledge, patterns, and solved issues across different users, sessions, and machines.
How it works
Hivemind intercepts agent interactions—including prompts, tool calls, and responses—and stores them as structured traces in Deeplake. A background worker analyzes these traces to identify repeated patterns and codifies them into reusable SKILL.md files. These skills and traces are then made available to all connected agents via hybrid lexical and semantic retrieval, allowing agents to proactively recall relevant information during new sessions.
Who it’s for
Software engineering teams using AI coding assistants such as Claude Code, Cursor, OpenClaw, Codex, Hermes, or pi, who want to synchronize their agents' collective intelligence.
Highlights
- Cross-Agent Memory: Shares knowledge across different AI assistants and teammates in real time.
- Automated Skill Extraction: Mines session traces to automatically create reusable skill documents.
- Multi-Assistant Support: Integrates with a wide range of tools including Claude Code, Cursor, and an MCP server for Claude Desktop.
- Cost Efficiency: Reduces token usage and the number of turns required to reach an answer by providing prior work in scope at recall time.
- Privacy Controls: Supports per-directory configuration to opt out of data collection for sensitive repositories.