MCP.Market

ProductBrain MCP server

by moxzas·io.github.moxzas/productbrain·v0.1.0
No reviews yet

Drive your ProductBrain product plan (goals, needs, bets, jobs) from any MCP client.

B83/100grade B
Write a reviewInstall
What users say
No reviews yet
Be the first
Safety scan
B83/100

full report

Adoption
Growing

0 stars225 downloads/wk

Reviews

Write one

Nobody has reviewed ProductBrain yet.

If you have run it, two minutes of your experience saves the next person an afternoon.

ProductBrain tools (27, 9 write)

write = sends, deletes, buys or posts

Read from the package source without running it. The installed server may list more.

  • add_agent_seat

    Mint an agent seat on a project (owner only, Team tier). Returns a pb_ API key ONCE — store it now, there is no retrieval.

  • archive_project

    Archive (archived:true) or unarchive (false) a project you own. Archiving hides it from list_projects without deleting anything — reversible, unlike delete.

  • create_projectwrite action

    Create a project you own and seed its system "Later" phase, so nodes can be added immediately. id defaults to a slug of the name.

  • create_share_linkwrite action

    Mint a PUBLIC read-only link to the project view. Pass expiresInDays for a TTL (otherwise it never expires) and settings (e.g. { includedGoalIds: [...] }) to scope it. Treat the returned token as a secret.

  • create_webhookwrite action

    Register an outbound webhook. events defaults to all (node.added|updated|deleted, iteration.added|updated|deleted). Returns the signing secret ONCE — store it immediately.

  • delete_webhookwrite action

    Remove a webhook registration. Irreversible — the secret is gone; re-registering mints a new one.

  • export_okf

    Serialize the whole plan to a portable OKF file bundle (markdown with frontmatter, one file per node). Use when you want the plan as files — for a repo, a diff, or handing it to a tool that reads markdown — rather than as JSON.

  • get_changelog

    Read the plan's history: every add/update/delete/restore, newest first, each attributed to the app or to a named agent. This is how you find a node AFTER it was deleted — /nodes can no longer see it, but restore needs its id. Filter by nodeId, action (add|update|delete|restore), source (app|api_key), agentId, since/before (ISO). Page with before=<_meta.next_before>, never an offset. Then undo with

  • get_tree

    Get a node in context. relation = context (ancestors+siblings+children, default) | ancestors | subtree | siblings. The endpoint's own names (up | down | lateral | context) are accepted too. There is no children-only view: subtree/children both return the full subtree.

  • invite_contributor

    Invite a guest contributor by email to one or more projects you own (owner only, Team tier). Sends a passwordless Clerk invite; one seat regardless of project count.

  • list_iterations

    List phases/iterations. current:true returns the active phase; status:active filters to active.

  • list_members

    List the members (human and agent seats) on a project, plus the owner. Any member can read this.

  • list_nodes

    List nodes in a project (bulk read). Optional type (goal|need|approach|job|task) and iteration filters.

  • list_projects

    List the projects you own or are a member of. Archived ones are hidden unless includeArchived is true.

  • list_share_links

    List every share token minted for a project, with createdAt/expiresAt/revokedAt and an `active` flag. The audit half of the share loop.

  • list_webhooks

    List the outbound webhooks registered on a project (secrets are never returned). Shows lastStatus/lastDeliveryAt so you can spot a failing receiver.

  • mutatewrite action

    Create/update/delete plan nodes and phases. action = add|update|delete|batch|addIteration|updateIteration|deleteIteration|classify. Put the rest of the v1 mutate body in payload (node | nodeId+data | mutations | iteration). Pass idempotencyKey (a UUID) to make a retry safe — a replay returns the original result instead of double-creating.

  • read_view

    Read what the user is looking at now: view_mode, selected_iteration, expanded_nodes, focused_goals, selected_node.

  • rename_project

    Change a project's display name (owner only). The id is immutable — keep using the same projectId afterwards.

  • revoke_memberwrite action

    Remove a member (human or agent) from a project. Removing an agent seat revokes its API key once its last membership is gone.

  • revoke_share_link

    Revoke a share token. The public link stops serving within ~30s (edge cache); the record dies immediately.

  • run_workflowwrite action

    Run the Story-Mapping methodology workflow (flow = add | task-curate | phase-assign). Returns the next prompt + opaque state, or a terminal changeset with a focus hint. Put the v1 workflow body in body.

  • search

    Semantic search over the ProductBrain plan. Returns nodes ranked by similarity with ancestor paths. Use this FIRST for any lookup — never scan all nodes by label.

  • set_view

    Drive the live canvas. command = { type: setViewMode|selectIteration|zoomToNode|expandNode|collapseNode|focusGoals, ... }.

  • submit_feedbackwrite action

    Tell the ProductBrain team how PB should improve: a feature you needed, friction you hit, a bug, or a general note. Stored for the team to read, not added to any plan. Make the summary self-explanatory (there is no reply channel). Keep credentials, customer records and confidential plan content out. Max 20 per 24h, so batch related points.

  • tier_status

    Read your own tier, API access, AI-credit balance, seat/project/node limits and billing period. Check this BEFORE a large run so you self-throttle instead of discovering a 402 mid-flight.

  • update_webhookwrite action

    Reconfigure a webhook in place (same id). Change url and/or events without minting a new secret, or pass rotateSecret:true to mint a new one (returned once). Prefer this over delete+recreate, which breaks the receiver.

Public scan report

scanner v0.1.10 · 2026-10-09 · same rubric, same numbers if you re-run it

no findings
  • Code scan3 source files scanned25/25
  • –Live reliabilityno gateway calls yet and no remote to proben/a
  • –Tool poisoningtools not inspected (local package is not executed); not countedn/a
  • Auth qualitystatic API keys via environment variables6/15
  • Maintenancelast push 6 days ago15/15
  • Maintainer identityregistry namespace matches repository owner; GitHub account older than a year8/10
Overall 83/100. Components that don't apply are left out of the denominator. Any critical finding is an F.RubricAppeal a findingJSON

How ProductBrain compares

Against every graded E-commerce MCP server, from our own scans and uptime checks. Updated after each scan run.

Safety score
83/100
higher than 77% of 2,174 E-commerce servers
Tools
27
the typical E-commerce server has 6
Tools that change things
33%
typical: 0%. Fewer write tools means less an agent can break
Downloads a week
225
more than 79% of 411 E-commerce servers with a package

What its tools let an agent do: 12 read · 4 create · 5 change · 4 delete · 1 send or post · 1 run or deploy

Install ProductBrain in Claude Code, Cursor or VS Code

Runs npx -y @productbrain-com/mcp on your machine. Read the scan report first; the gateway never runs local packages.

claude mcp add productbrain -- npx -y @productbrain-com/mcp
Add to Cursor

What the publisher says

From the ProductBrain repository's README, as published. We do not edit it. Read it on GitHub

ProductBrain MCP server

Drive your ProductBrain plan from an MCP client — Claude Desktop, Cursor, Claude Code — with tools for search, read, mutate, the methodology workflow, and the live view.

Design — a thin shim, on purpose

This server is a thin transport over ProductBrain's versioned v1 REST API. Each tool is one call to an existing /api/v1 endpoint. The REST contract stays canonical:

  • It's stable when the MCP spec churns — the contract you depend on is the frozen v1 API, not the protocol.
  • You can drop to raw HTTP or bring your own LLM at any time; MCP is one front-door, not the only one.
  • Responses carry the same in-band _meta coaching the API returns, so your agent self-corrects.

Install

// Claude Desktop — claude_desktop_config.json
{
  "mcpServers": {
    "productbrain": {
      "command": "npx",
      "args": ["-y", "@productbrain-com/mcp"],
      "env": {
        "PRODUCTBRAIN_API_KEY": "pb_your_key",
        "PRODUCTBRAIN_PROJECT_ID": "your-project-id"
      }
    }
  }
}
# Claude Code
claude mcp add productbrain \
  -e PRODUCTBRAIN_API_KEY=pb_your_key \
  -e PRODUCTBRAIN_PROJECT_ID=your-project-id \
  -- npx -y @productbrain-com/mcp

Cursor and other clients: add an mcpServers entry with the same command/args/env.

Config (env)

Tools

Every agent-facing /api/v1 endpoint has a tool. One tool, one endpoint, same _meta coaching passed back verbatim.

The plan

The view

Projects

Webhooks

Share links

Shortened. The full README is on GitHub.

Nothing above is checked by us. What we check is on the safety report.

ProductBrain: common questions

Is ProductBrain MCP server safe?
Mostly: it is graded B (83/100). Read the ProductBrain safety report
How do I install ProductBrain?
It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.
Does ProductBrain need an API key?
Yes. The registry entry asks for PRODUCTBRAIN_API_KEY.
Is ProductBrain maintained?
The last commit was 31 days ago (2026-10-04). The latest release is v0.1.0.
What can I use instead of ProductBrain?
Servers from other publishers that do the same job: PostHog MCP Server, Loki Mode MCP server and New-Term Radar MCP server.
How does ProductBrain compare with other E-commerce MCP servers?
safety score 83/100 (higher than 77% of 2,174 E-commerce servers); tools 27 (the typical E-commerce server has 6); tools that change things 33% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.

Alternatives to ProductBrain

Same job from other publishers: the closest match first, then the best rated.

  • PostHog MCP Server
    Official PostHog MCP Server for product analytics, feature flags, experiments, and more.
    A
  • Loki Mode
    Autonomous spec-to-product coding-agent CLI with an MCP server exposing 39 tools over stdio.
    B
  • New-Term Radar
    New AI model, product and repo names as they surface, and the search terms behind them, sized.
    B
  • Idea Reality
    Pre-build reality check. Scans GitHub, HN, npm, PyPI, Product Hunt — returns 0-100 signal.
    B
  • Stripe
    MCP server integrating with Stripe - tools for customers, products, payments, and more.
    A