← Back to Discover

Data Olympus

A git-based knowledge base format with a single-writer MCP server and CLI, meant to keep coding agents aligned with the standards and decisions a team has already accepted.

MCP serversNotes & knowledge
Visit website ↗

What it does

Data Olympus is a format and server for keeping project knowledge, such as engineering standards, architectural decisions and team guidance, as a directory of markdown files in git. Documents carry a stable id and controlled type, status and tier fields, and a supersedes chain lets you trace how a decision changed over time. It builds on Google's Open Knowledge Format v0.2, so tools that read that format can read it too. Agents reach the knowledge base through an MCP server that is the only writer. Proposed edits wait in a pending queue before they are committed, and the server uses advisory locks, per-session worktrees and a durable push queue to avoid write races. You can filter documents by status, tier and type. The project describes its job as pointing an agent to the rule that should govern a choice, such as a library, a pattern or a migration. It says it is not a code search or find-usages tool. The repository labels itself a pre-1.0 beta and pre-release, with stable releases distributed through PyPI and GHCR.

Capabilities

  • Knowledge base kept as plain markdown with YAML frontmatter in git
  • Single-writer MCP server and a CLI
  • Pending queue for proposed edits before they are committed
  • Filter by status, tier or type, and trace decisions through supersedes chains
  • Compatible with Open Knowledge Format v0.2 consumers
  • Stable releases on PyPI and GHCR

When to use it

  • Your team wants coding agents to follow decisions and standards you have already agreed on
  • You want a version-controlled, reviewable knowledge base with no proprietary service or database
  • Several agents write to the same knowledge base and need protection from write conflicts