Memory with reasons,
for coding agents.
Ergo is a memory backend for AI coding agents that stores every decision alongside why it was made — with kill-criteria, conflict detection, and full revision history. Stop losing context across sessions.
What makes it different
Most agent memory systems store text. Ergo stores decisions, the reasoning behind them, and what would make them wrong.
Reasons, not just claims
Every stored claim carries the reasoning that produced it. Retrieve the why, not just the what.
Kill-criteria built in
Each claim declares conditions under which it becomes false. Ergo checks them and flags stale beliefs before they mislead.
Conflict detection with NLI
Semantic contradiction detection catches when a new claim contradicts an existing one — flag, warn, block, or supersede.
Per-agent, per-repo scope
Memory scoped to the project + agent that created it. No cross-contamination between codebases or coding assistants.
Full revision history
Nothing is ever silently overwritten. Every change is a superseded event with a pointer back to what came before.
MCP tools, zero-config
Drop-in Model Context Protocol server. Works with Claude Code, Opencode, and anything that speaks MCP.
Status
Ergo is under active development. It's currently being dogfooded by the coding agents building it (Claude Code and Opencode) — the memory backend they use is Ergo itself.
- Core memory API + MCP toolsLive
- NLI-based conflict detectionLive
- Kill-criteria engineLive
- Team server MVP (multi-user)In progress
- SaaS / hosted offeringPlanned