Get started

Know if an idea is worth building before you ship. Preuve gives Cursor, Claude, or Codex a viability score, competitors, and source-linked evidence, the same analysis as preuve.ai. Paste one prompt, or run one command. Starter scans are free.

Get access

Two doors, and only one of them puts a key in your hands.Connector (OAuth). Add https://mcp.preuve.ai/mcp as a connector in claude.ai, or in any MCP client that signs you in over OAuth. This works on a free account: Preuve mints the key and installs it in the client for you, so there is nothing to copy or store. Starter scans cost nothing, a deep scan still spends paid quota. Skip straight to the claude.ai tab in step 2, and revoke the connection anytime from API key settings.Manual key. The header-based setups below (Claude Code, Cursor config, Codex, mcp-remote) and direct HTTP calls need a key you hold yourself. Create one at Account → API Keys, self-serve on Founder and above. It is the only path that shows you a prv_... value, so copy it when it appears, it’s shown once.

export PREUVE_API_KEY=prv_...

Connect

One connection, then your agent can score ideas in the chat. The claude.ai tab is the no-key path. Every other tab sends your prv_... key as a header, so keep it out of the chat.

Paste this as-is once you have a manually created key. The agent will tell you where to put it.

I want to validate startup ideas with Preuve: viability score, competitors, source-linked evidence.

Add the Preuve MCP server.

1. I have a Preuve API key (prv_...) from https://preuve.ai/app?settings=apiKeys.
   Do NOT ask me to paste the key into chat. Put it in MCP config / env.

2. Add the remote server at https://mcp.preuve.ai/mcp (HTTP):
   - Claude Code:
     claude mcp add preuve --transport http https://mcp.preuve.ai/mcp \
       --header "Authorization: Bearer <key>"
   - Cursor / Claude Desktop: ~/.cursor/mcp.json (or Claude Desktop config):
     {
       "mcpServers": {
         "preuve": {
           "url": "https://mcp.preuve.ai/mcp",
           "headers": { "Authorization": "Bearer <key>" }
         }
       }
     }
   - Codex:
     export PREUVE_API_KEY=<key>
     codex mcp add preuve --url https://mcp.preuve.ai/mcp \
       --bearer-token-env-var PREUVE_API_KEY
   - Clients that only support stdio:
     npx -y mcp-remote https://mcp.preuve.ai/mcp \
       --header "Authorization: Bearer <key>"

3. Confirm these 8 tools appear: start_analysis, get_analysis,
   enrich_analysis, export_analysis, create_batch, get_batch, export_batch,
   generate_ideas. There are also two prompts, validate_idea and
   generate_ideas, that you can pick from the client.

4. First test with scanType "starter" only (it costs nothing):
   start_analysis on my idea, poll get_analysis, show me the viability score.

Read https://docs.preuve.ai/agent-skill.md first.

claude mcp add preuve --transport http https://mcp.preuve.ai/mcp \
  --header "Authorization: Bearer $PREUVE_API_KEY"

Add this to ~/.cursor/mcp.json (or the project .cursor/mcp.json):

{
  "mcpServers": {
    "preuve": {
      "url": "https://mcp.preuve.ai/mcp",
      "headers": {
        "Authorization": "Bearer prv_..."
      }
    }
  }
}

Save the key in PREUVE_API_KEY, then:

codex mcp add preuve \
  --url https://mcp.preuve.ai/mcp \
  --bearer-token-env-var PREUVE_API_KEY

In claude.ai (web or desktop): Settings → Connectors → Add custom connector. Paste:

https://mcp.preuve.ai/mcp

Sign in with Preuve and approve. Works on a free account. There is no key to copy: Preuve mints one for the connector and installs it, and the value is never shown. Revoke it anytime from API key settings.

If your client cannot use a remote HTTP MCP server, bridge with mcp-remote:

{
  "mcpServers": {
    "preuve": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.preuve.ai/mcp",
        "--header",
        "Authorization: Bearer prv_..."
      ]
    }
  }
}

Tool list and extra install options: MCP server.

Get a score

Ask. A starter scan is free and usually done in 1-3 minutes.

Validate this idea with a starter scan: "AI bookkeeping for solo lawyers in France"

A deep scan spends your paid quota. Ask for one only when you want the full report.

Next steps

MCP tools

Score, enrich, and export an idea from the chat.

Agent Skill

Same validation, without your agent spending a paid scan by accident.

HTTP API

Same verdict, as JSON, from your own code.

Quotas

Starter is free. Deep uses the same quota as the web app.

Getting Started

For agents

Concepts

Next steps

MCP tools

Agent Skill

HTTP API

Quotas

​Next steps

MCP tools

Agent Skill

HTTP API

Quotas

Next steps