Skip to main content
Glama

Noggin

Noggin is a local-first brain for humans, teams, and AI agents. It ingests surface activity from Slack, agent sessions, GitHub, and direct CLI input; sends the content to Noggin Workers, an LLM brain agent; stores provenance in SQLite; and exposes the brain through CLI, MCP, Hermes, OpenClaw, a Slack slash command, and a small dashboard.

The product goal is simple: every useful mistake, decision, process detail, and workflow lesson should become reusable context instead of disappearing at the end of a chat.

V1 Surface Map

Slack / GitHub / Agent / CLI
          │
          ▼
  source event envelope
          │
          ▼
 validate -> redact -> dedupe -> Noggin Workers
          │                    │
          ▼                    ▼
 append-only event log   observations + entities + edges
          │                    │
          │                    └────────► Markdown graph nodes
          └────────► local SQLite brain ◄──────── skill proposals
                                │
                                ▼
             CLI + MCP + Hermes + OpenClaw + dashboard

Related MCP server: waypath

Install

Guided setup:

git clone <this-repo>
cd noggin
./setup.sh

The setup script creates .venv, installs noggin, writes a private ~/.noggin/noggin.env, asks for your LLM provider/API key, and walks through optional Slack, GitHub, sync, Hermes, and OpenClaw configuration.

Manual setup:

git clone <this-repo>
cd noggin
python3 -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"
noggin doctor

Noggin is LLM-only. Running the CLI, servers, and adapters requires an API key for the configured provider:

export NOGGIN_PROVIDER=openai
export NOGGIN_API_KEY=...
export NOGGIN_MODEL=gpt-4o-mini
noggin doctor

Supported providers:

Provider

API key env

Notes

openai

NOGGIN_API_KEY or NOGGIN_OPENAI_API_KEY

OpenAI chat completions

anthropic

NOGGIN_API_KEY or NOGGIN_ANTHROPIC_API_KEY

Anthropic Messages API

gemini

NOGGIN_API_KEY or NOGGIN_GEMINI_API_KEY

Gemini generateContent

openrouter

NOGGIN_API_KEY or NOGGIN_OPENROUTER_API_KEY

OpenAI-compatible

groq

NOGGIN_API_KEY or NOGGIN_GROQ_API_KEY

OpenAI-compatible

together

NOGGIN_API_KEY or NOGGIN_TOGETHER_API_KEY

OpenAI-compatible

mistral

NOGGIN_API_KEY or NOGGIN_MISTRAL_API_KEY

OpenAI-compatible

ollama

NOGGIN_API_KEY or NOGGIN_OLLAMA_API_KEY

OpenAI-compatible local server

custom

NOGGIN_API_KEY

Set NOGGIN_BASE_URL

Optional settings: NOGGIN_BASE_URL, NOGGIN_MODEL, NOGGIN_LLM_TIMEOUT, NOGGIN_TEMPERATURE, and NOGGIN_GRAPH_DIR.

Quick Start

export NOGGIN_PROVIDER=openai
export NOGGIN_API_KEY=...
noggin ingest "Decision: we keep the first version local-first and use MCP as the host-neutral adapter."
noggin ingest --source agent --kind mistake "Mistake: auto-editing skills silently breaks trust. Always create a proposal first."
noggin recall "local-first adapter"
noggin graph sync
noggin skills propose --content "Mistake: deployment failed because migrations were not listed in the release checklist."
noggin dashboard --open

Knowledge Arrangement

Noggin Workers arrange knowledge as source-backed triples:

source event -> observation -> subject --predicate--> object

The LLM chooses durable observations, subjects, predicates, objects, confidence, and tags. Noggin stores the event log, observations, entities, and edges in SQLite, then materializes each entity as a Markdown node file under ~/.noggin/graph/nodes/ by default. Node files include observations, outgoing links, incoming backlinks, and event provenance.

Graph commands:

noggin graph sync
noggin graph list
noggin graph show "noggin"

Adapters

Local Install For Hermes

scripts/install-hermes.sh
hermes plugins enable noggin

Local Install For OpenClaw

scripts/install-openclaw.sh

MCP

noggin mcp

Expose this command as a stdio MCP server. It supports:

  • brain_ingest

  • brain_recall

  • brain_reflect

  • brain_skill_propose

  • brain_graph_sync

  • brain_graph_list

  • brain_graph_show

Slack

export NOGGIN_SLACK_SIGNING_SECRET=...
noggin slack serve --port 8787

Configure a Slack slash command to POST to /slack/command.

Supported slash command text:

  • remember <text>

  • recall <query>

  • propose-skill <mistake or workflow lesson>

  • status

GitHub

noggin github issue owner/repo 123
noggin github pr owner/repo 456

Set GITHUB_TOKEN for private repositories or higher rate limits.

Hermes and OpenClaw

Install the generated skill files from integrations/hermes/SKILL.md or integrations/openclaw/SKILL.md. Hermes can also load the plugin in integrations/hermes/noggin_plugin/.

Skill Proposal Safety

Noggin Workers can draft skill changes, but the brain never silently edits skills by default. It creates a proposal with provenance and an explicit target path. Applying a proposal requires an allowed root and writes an audit event. If --run-tests is provided, the proposal is rolled back when tests fail.

Storage

Default database path:

~/.noggin/brain.db

Override with:

export NOGGIN_DB=/path/to/brain.db
export NOGGIN_GRAPH_DIR=/path/to/graph

Development

python -m pytest
ruff check .
A
license - permissive license
-
quality - not tested
B
maintenance

Maintenance

Maintainers
Response time
Release cycle
Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Servers

  • A
    license
    -
    quality
    B
    maintenance
    Local-first persistent memory layer for AI agents. Provides hybrid search (FTS5 keyword + vector embeddings) over 223K+ knowledge chunks via MCP. Tools: brain_search, brain_store, brain_entity, brain_subscribe. Features pub/sub with stable agent identity, delivery tracking, and Claude --channels integration. SQLite + BrainBar Swift daemon on Unix socket.
    Last updated
    8
    Apache 2.0
  • A
    license
    A
    quality
    B
    maintenance
    Local-first external brain for Claude Code, Codex, and any MCP client. Stores decisions, entities, and session artifacts in one SQLite file and exposes MCP tools for recall, page, promote, review, graph-query, and source-status.
    Last updated
    11
    9
    4
    MIT
  • A
    license
    -
    quality
    C
    maintenance
    Local-first knowledge backend for AI agents that connects MCP hosts to an Obsidian-compatible vault with indexed retrieval, token-budgeted memory recall, and secure ingestion.
    Last updated
    1
    MIT

View all related MCP servers

Related MCP Connectors

  • User-owned memory for AI agents, Copilot, Claude, IDEs, CLIs, and chat apps over remote MCP.

  • Private-by-default, local-first memory/context/task orchestrator for MCP apps and agents.

  • Local-first RAG engine with MCP server for AI agent integration.

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Edward-Zion-Saji/noggin'

If you have feedback or need assistance with the MCP directory API, please join our Discord server