Nauro

A real Codex session, then a Claude session in Pareto, Nauro's development mock project.

Pennykeep, the bundled demo, needs no account or agent setup:

mkdir -p /tmp/nauro-demo && cd /tmp/nauro-demo nauro init --demo nauro check-decision "Store dollar amounts as decimal numbers"

The top result isD001,Amounts stored in integer cents, never floating point. Its rationale explains why floating point makes money totals drift.

cd your-repo nauro adopt --with-skills --with-subagents

Restart your agent, then invoke/nauro-adoptin Claude Code or$nauro-adoptin Codex. The skill reads the repo, asks about missing rationale, and seeds the project record. The optional flags also install the gatednauro-ship-taskworkflow and its four workflow agents.

Runnauro statusto check MCP, skills, and workflow agents. Runnauro doctorto check the project store.

Nauro surfaces prior judgment for the agent to assess. Retrieval is advisory. New or revised judgment requires your explicit approval. Local use needs no account, and the record is plain Markdown on your machine. Cloud sync and remote MCP access are optional. Nauro sends no product analytics.

Nauro preserves an existingAGENTS.mdunless you explicitly runnauro sync. A# Manualsection survives replacement.

A free, read-only viewer for the project record: timeline, map, list, activity, and docs. First launch installs the CLI, connects your agents, and adopts or attaches a project.

If a small repo plus a reliable AGENTS.md or CLAUDE.md keeps agents oriented, Nauro may be more than you need.

Nauro is intended for longer-lived work where context can decay across sessions, tools, repositories, machines, or repeated handoffs.

uv sync --all-packages --all-extras uv run pytest packages/nauro-core/tests/ -x -q uv run pytest packages/nauro/tests/ -x -q

Apache 2.0. Named for Peter Naur'sProgramming as Theory Building(1985).

Atlaso is a memory layer for AI tools — connect once and it automatically recalls the context that matters across Claude Code, Cursor, Codex, and every AI tool you use.

Privacy-first local memory vault every AI shares over MCP. Markdown + SQLite on your machine; Claude, ChatGPT, Cursor, and any MCP client read and write it live. No cloud, no account, no telemetry.

Cross-vendor AI memory over MCP. One semantic memory store, readable and writeable from every MCP client — Claude, ChatGPT, Cursor, Windsurf, Zed, Kimi.

Generate random word pairs for the Imposter party game directly via Claude or Cursor. Let AI host your next social deduction game!

Shared company context and governed tools for ChatGPT, Claude, Codex and other MCP clients.

You decide what Claude remembers about you. LocalKeep is a personal timeline you curate—titled, dated, editable banks on your machine—then Claude looks them up via MCP instead of inventing uneditable cloud memory.

Persistent AI memory shared across Claude, ChatGPT, coding agents, and compatible MCP clients.

Private persistent memory for Claude, ChatGPT & Gemini via MCP — semantic search, zero-code setup.

Turn LinkedIn saved posts into a queryable knowledge base you can search, chat with, and connect to Claude or ChatGPT through MCP.