nauro 0.3.0

Local CLI + MCP server: set your project's doctrine once, every connected AI agent inherits it.

Nauro

Set the doctrine once. Every agent inherits it.

Your project's doctrine — goals, decisions, rejected paths — is inherited by every connected agent. When an agent proposes an approach that conflicts with a past decision, Nauro catches the drift before it ships. Works with Claude, Perplexity, ChatGPT, Cursor, and any MCP client.

Install

pipx install nauro   # or: pip install nauro

Requires Python 3.10+.

Quickstart

nauro init --demo
nauro setup claude-code   # writes the MCP entry to ~/.claude/settings.json

Open Claude Code and ask:

"Check if we should add a WebSocket endpoint for live task updates"

The demo creates a sample project with 7 decisions, project state, and open questions. check_decision surfaces a conflict: the team already chose SSE over WebSocket because persistent connections weren't released during ECS rolling deploys. No account needed.

For real-project setup, cross-surface access, MCP tool reference, and architecture details, see the main project README.

Why Nauro?

Memory tools record what agents saw and said. Nauro captures what you decided and rejected, then checks every session against those decisions before they drift.

The check_decision → propose_decision → confirm_decision pipeline surfaces conflicts for you to confirm before they're written, across any connected surface. Decisions made in Claude Code surface in Perplexity. No platform vendor owns your context.

Pricing

Free: unlimited local usage, unlimited projects, 5,000 remote MCP calls/month. See nauro.ai/pricing for hosted tiers.

---

Apache 2.0 license. Part of the nauro-ai/nauro monorepo.

Named for Peter Naur, whose 1985 paper Programming as Theory Building argued the real program is the theory in the programmer's mind, not the code. Every fresh agent session is the equivalent of losing that programmer.