Model Context Protocol · Free

Put a GTM operations team inside your AI.

Connect LeanScale's delivery playbooks, proprietary field studies and 95 episodes of operator interviews straight to Claude, Cursor, or any MCP client. Ask it a real go-to-market question and get a cited answer from people who do this for a living — while you're doing the work, not after.

13Delivery playbooks
95Podcast episodes
7Field studies
1,857Searchable passages
The idea

Every consultancy writes a playbook nobody reads. This one answers questions at the moment you have them.

What's inside

Four bodies of work, one endpoint.

Not a marketing site scraped into a vector store. These are the actual delivery methodologies LeanScale's architects run on customer engagements, plus original research built from their own delivery data.

Playbooks

13 delivery playbooks

The methodology, structured Blueprint → Build → Enable → Maintain. GTM Lifecycle, Market Map, Sales Territory Design, Lead Routing, Growth Model, Quote to Cash, CPQ, Attribution, Onboarding, CRM Migration, and the two automation playbooks.

Playbook library →
Research

7 field studies

Built from LeanScale's own delivery logs, call transcripts and shared channels across a panel of B2B software companies — not from a survey. GTM tech stacks, AI adoption in GTM, RevOps investment ratios, and the Series B/C fundraise clocks.

Example study →
Interviews

95 podcast episodes

The LeanScale Podcast, fully structured — frameworks, verbatim quotes with timestamps, practical advice broken out by role, and the numbers guests actually cited. CROs, VPs of RevOps, founders.

Knowledge hub →
Case studies

32 customer case studies, anonymized

The problem a company had and how it was solved. Customer names, people, third parties and delivery provenance are stripped out at build time and [withheld] marks each removal — outcomes, metrics and quotes are held back entirely. What's left is the shape of the work, which is the part that transfers.

Named versions are internal-only
The tools

Eight tools your model can call.

Every response comes back with a source URL attached, so the model can cite where an answer came from instead of blending it into its own priors.

diagnose_gtm

Describe the mess, get the motion.

The one to reach for when someone describes a problem rather than asks about a topic. Takes a plain-language situation and returns which delivery motion applies, what it matched on, the relevant sections of the playbook, and supporting evidence from the studies and podcast.

search_leanscale

Search everything at once. Returns cited excerpts.

get_playbook

A full playbook, or one named section of it.

list_playbooks

Every playbook, by category, with its section headings.

get_benchmark

Proprietary benchmark data when the question is "what's normal".

search_podcast

Operator opinions, frameworks and quotes with timestamps.

about_leanscale

What LeanScale is and what this server can answer.

get_case_study

Anonymized customer case studies by sector or engagement type.

Try it

Run a query right here.

The same endpoint your AI client will call. Describe a go-to-market situation the way you'd actually say it out loud.

Uses a demo key. No signup needed to try.
Get access

Free, but not anonymous.

A work email gets you a key in about three seconds. We gate it so we know which teams are actually reaching for this — that's the whole reason it's free.

Personal addresses aren't accepted. Gmail, Outlook, iCloud and the rest are rejected — this is built for teams, and knowing the company is the point.

What we log. Your key is tied to your name, company and email. Every tool call records which tool ran and the query text. That's deliberate: it's how we know what people are actually trying to solve.

What we don't do. No mailing list, no drip sequence, no reselling. If something you ask looks like a problem we could help with, someone might reach out once.

Customer work is anonymized, not exposed. Case studies are stripped of names, people and third parties before they ever reach the corpus, and the build fails if any unrecognized proper noun survives. Named versions never leave a LeanScale key.

Install

Two minutes, whichever client you use.

Replace YOUR_KEY with the key from above. The endpoint is https://leanscale-mcp.netlify.app/mcp.

Claude Code

One command, from any terminal.

claude mcp add --transport http leanscale https://leanscale-mcp.netlify.app/mcp --header "Authorization: Bearer YOUR_KEY"

Claude Desktop / claude.ai

Settings → Connectors → Add custom connector. Paste the URL with your key on it — the connector UI has no header field, which is why the key also works as a query parameter.

https://leanscale-mcp.netlify.app/mcp?key=YOUR_KEY

Cursor / Windsurf / VS Code

Add to your mcp.json.

{
  "mcpServers": {
    "leanscale": {
      "url": "https://leanscale-mcp.netlify.app/mcp",
      "headers": { "Authorization": "Bearer YOUR_KEY" }
    }
  }
}

Anything else

Streamable HTTP transport, JSON-RPC 2.0. Bearer token in the Authorization header, or ?key= on the URL. A plain GET returns server info and the corpus stats without a key.

curl -s https://leanscale-mcp.netlify.app/mcp \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
Once it's connected

Just ask. "Our inbound leads take three days to get worked — what should we do?"