The hivemind for human-agent teams: shared memory, and the tools that put it to work.

Without shared team memory, every coding agent session starts cold. Agents repeat work already done by teammates, make implementation choices inconsistent with prior decisions, and have no awareness of what other agents on the team are doing in parallel.

SageOx captures context where your team actually works: meetings, Slack, and coding-agent sessions. Through the Ox CLI, SageOx continuously primes agents, so they have full team awareness instead of operating in isolation.

Every commit and pull request traces back to the decision or meeting that originated it.

→ sageox.ai

Software and product engineering teams that:

• Build primarily through prompts, using AI coding agents and assistants
• Run parallel agent sessions across multiple developers, repos, and machines
• Treat team context as core IP, not data exhaust

sageox/ox is our open source client. It connects your coding agent sessions to your team's memory.

brew tap sageox/tap
brew install ox

Setup and commands live in the ox README. Everything else is in the docs.

The CLI talks to the hosted SageOx platform. These are products, not repos:

• Ox MCP — the same team memory inside AI assistants, over Model Context Protocol
• Ox Dot — records in-person meetings and extracts the decisions
• Ox Console — search team memory, manage context, configure agent sessions

Get access at sageox.ai.