Cambium Product Manager MCP — Bring Synthetic Personas Into Your Claude Sessions
LocalAbout Cambium Product Manager MCP — Bring Synthetic Personas Into Your Claude Sessions
Cambium Product Manager MCP brings synthetic personas grounded in public demographic and market data into Claude for product research and decisions. It retains business context, project phase, and past decisions so work can continue across sessions.
Key features
- 01Guide setup and retain business context
- 02Assess ideas with persona interviews and web research
- 03Analyze demographics and regulatory risks
- 04Adapt guidance to discovery, build, or scale phases
- 05Track decisions and restore context across sessions
Use cases
- 01Evaluate a new product idea and its market
- 02Refine a product using target-audience reactions
- 03Review launch plans for competition and compliance risks
- 04Continue research and record decision rationales
MCP support
It connects to Claude as a remote MCP server, providing product research, business context, phase management, decision tracking, and session continuity. The README describes adding it over HTTP and authenticating an account.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMECambium AI - product manager MCP
Features
Business Case Setup
Define your company context, products, business model, strategic goals, constraints, and target market. The MCP guides you through a structured interview to capture everything Claude needs to give relevant product advice. Your business case persists across sessions.
AI Product Validation
Validate any product idea with AI-powered analysis that combines persona interviews, web research, demographic data, and regulatory assessment. Just describe an idea (e.g. "A smart water bottle that tracks hydration via built-in sensor") and get a decision-grade brief with a clear verdict — Strong Go, Conditional Go, Needs More Research, Pivot, or No Go. If you don't have an idea yet, one can be generated for you based on your business case.
Decision Tracking
Record product decisions (pivot, proceed, kill, scope change, feature prioritization) along with the rationale behind them. Decisions accumulate across sessions, giving you a living decision log you can reference later.
Project Phase Management
Track where you are in the product lifecycle - discovery, build, or scale. Phase-aware guidance adapts the research and recommendations to what matters most at your current stage.
Personas
Product research is powered by Cambium AI’s personas: synthetic demographic profiles grounded in public data sources such as the U.S. Census. Each persona is built from real demographic attributes - such as age, location, income, education, and family status - and then supplemented with AI-generated layers including biography, personality, goals, frustrations, motivations, and media habits. This means the personas are anchored in observed population characteristics, while the richer narrative context is model-generated. The result is a research tool designed to help teams explore responses through the lens of data-grounded audience profiles, rather than relying on assumed or purely invented segments.
Session Continuity
The MCP remembers your context between sessions. It recalls your business case, past decisions, project phase, and a summary of your last session so you can pick up right where you left off.
Sign up for Cambium AI
If you don't have an account yet, head to https://app.cambium.ai/ and sign up before continuing.
Add the Cambium AI - product manager MCP
Add the MCP to your Claude session:
claude mcp add --transport http cambium-ai--product-manager https://api.app.cambium.ai/mcp/product-manager/mcp
You should see the output:
josh:test-product-manager-mcp-prod$ claude mcp add --transport http cambium-ai--product-manager https://api.app.cambium.ai/mcp/product-manager/mcp
Added HTTP MCP server cambium-ai--product-manager with URL: https://api.app.cambium.ai/mcp/product-manager/mcp to local config
File modified: /home/joshgillott/.claude.json [project: /home/joshgillott/cambium/test/test-product-manager-mcp-prod]
Launch Claude Code
Launch Claude Code:
claude
Should enter a Claude code session (please forgive the spacing issues):
josh:test-product-manager-mcp-prod$ claude
╭─── Claude Code v2.1.85 ────────────────────────────────────────────────────────────────────────╮
│ │ Tips for getting started │
│ Welcome back Josh! │ Run /init to create a CLAUDE.md file wit… │
│ │ ───────────────────────────────────────── │
│ ▐▛███▜▌ │ Recent activity │
│ ▝▜█████▛▘ │ No recent activity │
│ ▘▘ ▝▝ │ │
│ │ │
│ Opus 4.6 (1M context) · Claude Max · Josh Gillott │ │
│ ~/cambium/test/test-product-manager-mcp-prod │ │
╰────────────────────────────────────────────────────────────────────────────────────────────────╯
/remote-control is active. Code in CLI or at
https://claude.ai/code/session_01JPbcrguKi4yHstW4dVGBiK
⎿ Please upgrade to the latest version of the Claude mobile app to see your Remote Control
sessions.
──────────────────────────────────────────────────────────────────────────────────────────────────
❯
──────────────────────────────────────────────────────────────────────────────────────────────────
? for shortcuts
Your Claude Code output may look different to mine. e.g. name, model, subscription, directory, /remote-control. The important thing is that it launches!
Authenticate the MCP server
/mcp
Hit enter.
You should be able to see:
──────────────────────────────────────────────────────────────────────────────────────────────────
Manage MCP servers
4 servers
Local MCPs (/home/joshgillott/.claude.json [project:
/home/joshgillott/cambium/test/test-product-manager-mcp-prod])
❯ cambium-ai--product-manager · △ needs authentication
Hit enter.
❯ /mcp
╭────────────────────────────────────────────────────────────────────────────────────────────────╮
│ Cambium-ai--product-manager MCP Server │
│ │
│ Status: △ needs authentication │
│ Auth: ✘ not authenticated │
│ URL: https://api.app.cambium.ai/mcp/product-manager/mcp │
│ Config location: /home/joshgillott/.claude.json [project: │
│ /home/joshgillott/cambium/test/test-product-manager-mcp-prod] │
│ │
│ ❯ 1. Authenticate │
│ 2. Disable │
╰────────────────────────────────────────────────────────────────────────────────────────────────╯
Hit enter (select 1. Authenticate).
You will be redirected to https://app.cambium.ai/ to sign in.
Sign up or log in as you usually would. You will see the Authorize Access screen:
You should see the Authentication Successful screen and be redirected to Claude Code in your terminal.
Terminal output now acknowledges the connection:
❯ /mcp
⎿ Authentication successful. Connected to cambium-ai--product-manager.
──────────────────────────────────────────────────────────────────────────────────────────────────
❯
──────────────────────────────────────────────────────────────────────────────────────────────────
? for shortcuts
Possible unhappy path - hopefully this was an issue that has been resolved by recent updates
I had to restart Claude Code for the authentication to apply correctly:
josh:test-product-manager-mcp-prod$ claude
╭─── Claude Code v2.1.84 ────────────────────────────────────────────────────────────────────────────────────────────────╮
│ │ Tips for getting started │
│ Welcome back Josh! │ Run /init to create a CLAUDE.md file with instructions for Claude │
│ │ ───────────────────────────────────────────────────────────────── │
│ ▐▛███▜▌ │ Recent activity │
│ ▝▜█████▛▘ │ No recent activity │
│ ▘▘ ▝▝ │ │
│ │ │
│ Opus 4.6 (1M context) · Claude Max · Josh Gillott │ │
│ ~/cambium/test/test-product-manager-mcp-prod │ │
╰────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
/remote-control is active. Code in CLI or at
https://claude.ai/code/session_01ArYuFKwp2RTw9JevDD2b5D
❯ /mcp
────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
Manage MCP servers
4 servers
Local MCPs (/home/joshgillott/.claude.json [project: /home/joshgillott/cambium/test/test-product-manager-mcp-prod])
❯ Cambium-AI--product-manager · ✔ connected
claude.ai
claude.ai Atlassian · ✔ connected
claude.ai Gmail · △ needs authentication
claude.ai Google Calendar · △ needs authentication
https://code.claude.com/docs/en/mcp for help
↑↓ to navigate · Enter to confirm · Esc to cancel
Once you've added and authenticated the MCP, just talk to Claude naturally. Claude will call the right tools behind the scenes.
Research and validate a product idea
This is the core feature. Describe a product idea and Claude will research and validate it end-to-end:
"Research this idea: a smart water bottle that tracks hydration via a built-in sensor"
If you don't have an idea yet, Claude can generate one for you:
"Generate a product idea for me"
Claude will automatically ensure everything it needs is in place before running the research — collecting your business case, setting your project phase, and recalling prior context if this isn't your first session. You don't need to set these up separately.
Behind the scenes, four research agents work in parallel:
- Audience sizing — identifies your target demographic and selects real data-grounded personas
- Persona interviews — interviews those personas about your idea to surface reactions, objections, and willingness to pay
- Market research — searches the web for competitors, market size, and trends
- Regulatory analysis — flags compliance requirements and legal risks that could block launch
The result is a decision-grade brief with a clear verdict — Strong Go, Conditional Go, Needs More Research, Pivot, or No Go — backed by evidence from each agent. Research can take a few minutes. Claude will keep you updated on progress.
Business case
Before your first research run, Claude will walk you through a structured interview to capture your business context — company context, existing products, business model, strategic goals, constraints, and target market. This only happens once. Your business case persists across sessions.
You can also set it up or update it directly at any time:
"Help me set up my business case"
"Update my business case — we've pivoted to B2B"
"Show me my business case"
Project phase
Claude adapts its advice based on where you are in the product lifecycle:
- Discovery — stress-test assumptions, avoid premature commitment
- Build — ship a focused MVP, frame trade-offs
- Scale — find growth levers, defend competitive moat
Claude will ask about your phase during onboarding. You can change it any time:
"We're moving into build now — we validated the core assumption"
Record decisions
After reviewing research, tell Claude your decision and the rationale behind it. Decisions accumulate across sessions, giving you a living decision log.
"Let's proceed with this — the market validation is strong enough for an MVP"
"Kill this idea — the regulatory risk is too high for our timeline"
"Let's pivot — the original concept is too broad, but the hydration tracking angle has legs"
Session continuity
Claude automatically recalls your business case, past decisions, project phase, and a summary of your last session when you start a new conversation. It also saves a session summary when you disconnect. You can explicitly wrap up to capture specific open questions:
"Let's wrap up for today — next time I want to revisit the pricing model"
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.