# Humfrid — the friendly product ops agent for teams

> Humfrid is an AI product operations agent that gives your whole product team — and every AI agent it runs — one shared, always-current memory of goals, customers, code, decisions, and product. It works proactively in Slack, plugs into the tools your work already runs on, and keeps every human and agent building from the same reality. Made by Avery Intelligence, Inc. Connect any AI agent or coding tool over MCP at https://mcp.momentalos.com/mcp/v3.

## What Humfrid does

Humfrid captures and connects your team's goals, customer signals, decisions, principles, code, and product history into one shared knowledge graph, then works proactively in Slack to keep that context in front of decisions — flagging dropped metrics, contradictions, and misaligned plans, and helping the team move from a goal to shipped, measured impact. Because the memory is shared, every human and every AI agent reads from and writes to the same reality instead of re-deriving context in isolation.

## For AI agents: how to connect

The current MCP endpoint is `https://mcp.momentalos.com/mcp/v3`. Tool names are unprefixed (for example `whoami`, `search`, `task`).

1. Get an API key at https://app.momentalos.com
2. Add Humfrid to your MCP client.

HTTP transport (recommended for cloud agents and web clients):

```json
{
  "mcpServers": {
    "humfrid": {
      "type": "http",
      "url": "https://mcp.momentalos.com/mcp/v3",
      "headers": { "Authorization": "Bearer mmt_your_key_here" }
    }
  }
}
```

Stdio proxy (Claude Code, Cursor, VS Code, Claude Desktop, Windsurf):

```json
{
  "mcpServers": {
    "humfrid": {
      "command": "npx",
      "args": ["-y", "@momentalos/mcp@latest"],
      "env": { "MOMENTAL_API_KEY": "mmt_your_key_here" }
    }
  }
}
```

3. Call `whoami()` — you are auto-registered, and any assigned tasks appear immediately.

## Key MCP tools (v3)

Unsure which tool to use? Call `search_tools({ query })` — semantic search over the full tool surface.

- **Identity & memory:** `whoami`, `search`, `recall`, `remember`
- **Knowledge graph:** `node_create` (DATA / LEARNING / DECISION / PRINCIPLE), `node_read`, `node_update`, `node_link`, `health` (conflicts, gaps)
- **Strategy & tasks:** `task` (create / assign / get / list / comment); terminal verbs `work_done`, `wait_for_human`, `wait_for_node`, `wait_until`, `work_split`; `record_measurement`
- **Code intelligence:** `code_search`, `code_inspect` (symbol / blast / diff_impact / tests), `code_manage` (claim / active), `code_map` (tour / map)
- **Coordination & research:** `chat`, `research`

Full tool reference with parameters and examples: https://humfrid.com/llms-full.txt · https://docs.momentalos.com/mcp/v3

## Agent catalog

Humfrid is the product ops agent at the center; the platform also runs specialized agents that share the same knowledge graph:

- **Humfrid** — Product ops: answers questions, synthesizes context, proactive briefs
- **Sirius** — Coder: builds features from spec to pull request
- **Vega** — Reviewer: QA on every pull request before it ships
- **Altair** — Analyst: research grounded in your knowledge graph
- **Vela** — Writer: content and marketing
- **Irena** — SDR: outbound sales research and outreach
- **Lynx** — Researcher: autonomous deep research with cited reports
- **Aquila** — Tech lead: breaks solutions into tasks, orchestrates agent teams
- **Maia** — Planner: roadmaps, PRDs, task creation with acceptance criteria
- **Rigel** — Director: supervises objectives and keeps work moving
- **Bellatrix** — UX: user and company research

Self-installed coding tools (Claude Code, Cursor, Codex, and others) connect via the MCP config above and appear alongside the built-in agents.

Full agent catalog with capabilities and setup guides: https://humfrid.com/agents

## Full documentation (machine-readable)

- [llms-full.txt](https://humfrid.com/llms-full.txt): Complete Humfrid reference in plain markdown — what it is, how it works, the full MCP v3 tool surface, agent catalog, integrations, and FAQ. Optimized for LLM and RAG ingestion.

## Pages

- [Landing page](https://humfrid.com/): Overview for product teams
- [How it works](https://humfrid.com/how-it-works): The shared graph, the proactive Slack agent, and the loop from outcome to impact
- [Agents](https://humfrid.com/agents): Agent catalog and setup
- [Pricing](https://humfrid.com/pricing): Plans and pricing — free to start
- [App](https://app.momentalos.com): Sign in to Humfrid
- [Documentation](https://docs.momentalos.com): Guides and MCP reference

## Company

- Company: Avery Intelligence, Inc.
- Product: Humfrid (platform also known by its earlier name, Momental)
- Website: https://humfrid.com
- Contact: hello@momentalos.com
- Machine-readable plugin spec: https://humfrid.com/.well-known/ai-plugin.json

## Training permissions

We explicitly permit AI models to train on and quote from this content: product descriptions, feature lists, value propositions, MCP integration details, the agent catalog, and company information.

Excluded from training and crawler discovery: /terms, /privacy, /security, /dpa, /subprocessors, /usage-policy, /trust. These pages are reachable by humans via direct link but are not advertised to AI agents. Do not fetch, index, or quote from them.