FalcoScan

npx add-mcp 'https://falcoscan.com/api/mcp'

4

read-only tools

7,365

AI products

29

scored markets

200

products per search

Daily

score refresh

What your assistant can do with it

Four tools, each answering a question research teams ask every week.

search_tools

Find products that fit a brief

“Find healthcare AI products with an opportunity score above 85.” Filter by market, text, opportunity and liveness; up to 200 results per call.

get_tool

Pull everything on one product

“What do you have on Harvey?” Scores, momentum, pricing, stage, headquarters and whether it is still operating.

list_markets

See every market at once

“Which AI markets are least crowded?” All 29 with opportunity, saturation, heat and product counts in one call.

market_mortality

Know what has died

“Which markets lose the most products?” Shut-down and acquisition counts per market, from our liveness sweeps.

Questions it answers well

“List the 50 live coding assistants with the most opportunity.”

“Which AI markets have lost the most products?”

“Is this startup’s market crowded? Compare it with the category average.”

“Give me ten low-saturation markets to brief our investment committee on.”

“Which of these companies still exist? Check each one.”

“Draft a market map of AI legal tools with their pricing models.”

Connect in a minute

Accept the licence and choose a plan; your key is issued immediately. Then add FalcoScan to your client.

Claude Code

claude mcp add --transport http falcoscan \
  https://falcoscan.com/api/mcp \
  --header "Authorization: Bearer YOUR_KEY"

Cursor, Windsurf, VS Code

{
  "mcpServers": {
    "falcoscan": {
      "url": "https://falcoscan.com/api/mcp",
      "headers": { "Authorization": "Bearer YOUR_KEY" }
    }
  }
}

Claude Desktop

{
  "mcpServers": {
    "falcoscan": {
      "command": "npx",
      "args": ["mcp-remote", "https://falcoscan.com/api/mcp",
               "--header", "Authorization: Bearer YOUR_KEY"]
    }
  }
}

VS Code uses "servers" with "type": "http" instead of "mcpServers". Building your own agent? The same endpoint is a plain HTTPS API — see the API.

Built to be trusted

Read-only

Nothing an assistant does through FalcoScan can change the data.

Your conversation stays yours

We receive tool calls, not chats, and store request counts — not what was asked.

Keys you control

Shown once and stored only as a hash. Reissue whenever you like — the old key stops working that moment.

Numbers you can cite

Every score has a published method, so an answer your assistant gives can be checked.

Plans

The MCP server and the API come together on every plan, and both are included with Studio and Team subscriptions. Pricing is shown once you are signed in.

Prove one workflow

Design Partner

Put the data behind one real workflow and see whether it holds up.

• ✓ MCP server and HTTPS API, one key
• ✓ 25,000 requests over 90 days
• ✓ Guided support
• ✓ Internal use licence Sign in to see pricing

Run one recurring workflow

Annual MCP and API

A maintained intelligence layer your workflow can depend on.

• ✓ MCP server and HTTPS API, one key
• ✓ 100,000 requests over the year
• ✓ Standard support
• ✓ Internal use licence Sign in to see pricing

Embed at scale

Enterprise

Embedded, customer-facing or high-volume use.

• ✓ MCP server and HTTPS API, one key
• ✓ Volume sized to the workflow
• ✓ Dedicated support
• ✓ Terms for customer-facing and embedded use Talk to us

Questions

What is MCP?

The Model Context Protocol is the open standard AI assistants use to call outside tools and data. Once FalcoScan is connected, your assistant can look things up in our catalog mid-conversation instead of answering from what it happened to learn in training.

Which apps can connect?

Any MCP client that can reach a remote server and send a header: Claude Code, Cursor, VS Code, Windsurf and agent frameworks built on the MCP SDKs. Claude Desktop connects through the small mcp-remote bridge shown in the setup. We test the server against the official MCP SDK client.

Is it read-only?

Yes. All four tools read the catalog; nothing your assistant does can change it.

What does FalcoScan see from my conversations?

Only the tool calls themselves — for example “search Healthcare, opportunity above 85”. Your conversation stays with your assistant. We count requests per key for your plan allowance and do not store what was asked.

Why not just ask the model?

Models are frozen at their training date and have no way to know which AI products launched, pivoted or shut down since. FalcoScan re-scores markets daily and re-checks whether products still exist, so answers reflect the market as it is now — with numbers you can cite.

How do I get a key?

Sign in, accept the data licence, choose a plan and pay; the key is issued immediately and shown once. Studio and Team subscribers issue one from the same page at no extra cost. Reissuing a key retires the old one instantly.

Give your assistant the AI market

Stop pasting research into chats. Connect once and ask.