MCP Integration

Use Mera7 from any AI assistant

Evaluate any vendor directly from your AI assistant (Claude, ChatGPT, Cursor, Zed, any MCP client). Just ask “does X fit us?” and it sends your context privately to Mera7 for a personalized fit analysis. No forms, no meetings, no data shared with the vendor.

Just ask

“Does Acme fit us?” Claude already knows your company from your files and conversations. It structures your context automatically.

Private analysis

Mera7 matches your context against the vendor's verified product data. You get a personalized fit report. The vendor never sees your data.

Talk or schedule

Ask the vendor questions directly, or schedule a call. All from Claude. The vendor sees your question, never your raw context.

Install

Most MCP clients accept the same JSON config. Copy it below, then jump to your client for where to paste it.

MCP config

{
  "mcpServers": {
    "mera7": {
      "command": "npx",
      "args": ["-y", "mera7-mcp"]
    }
  }
}

Where to paste it

Claude Desktop

Paste into ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or the Windows/Linux equivalent. Restart the app.

Claude Code

One command: claude mcp add mera7 npx -y mera7-mcp

Cursor

Paste into ~/.cursor/mcp.json (global) or .cursor/mcp.json (per project).

Zed

Settings → Assistant → Context Servers. Paste the config block.

ChatGPT Desktop

Settings → Connectors → Add MCP server. Requires a plan with MCP connectors enabled. Point the server to 'npx -y mera7-mcp'.

Any other MCP client

Refer to your client's MCP config docs. Most accept the JSON above verbatim.

On first run, the MCP prints a verification URL + short code. Visit it, sign in, confirm. Token stored locally for 30 days.

Authentication

On first use, your MCP client will print a URL and a short code. Visit the URL, sign in to Mera7, and confirm the code to authorize the client.

Tokens are stored locally in ~/.mera7/mcp-token.json and last 30 days. You only do this once per machine.

Available tools

analyze_fit

Get a personalized fit analysis. Claude sends your context privately to Mera7, which matches it against the vendor's product data. Neither side sees the other's raw data.

Try: "Does Acme fit our company?"

list_vendors

Discover vendors available on Mera7

Try: "What vendors are on Mera7?"

query_vendor

Ask a targeted question about a vendor's product. The answer is grounded in the vendor's public knowledge base, not a full fit analysis.

Try: "Does Acme integrate with Salesforce?"

ask_vendor

Send a question directly to the vendor's team. They respond asynchronously.

Try: "Ask them about enterprise pricing"

check_questions

See if the vendor has responded to your questions

Try: "Did they answer my pricing question?"

Works with

Claude DesktopClaude CodeCursorZedAny MCP client

Privacy by design

Mera7 acts as neutral middleware. When you run a tool, your assistant sends a summary to Mera7 (not your raw files). Mera7 filters it through a 3-stage pipeline, then the vendor receives only a qualification brief. The vendor never sees your summary or raw context.

mera7.com · Privacy · Terms