PeakAI

Números de teléfono verificados y correos electrónicos laborales para prospectos B2B dentro de tu agente de IA. Busca personas y empresas por cargo, nivel de seniority, industria, ubicación y tamaño, encuentra contactos desde perfiles de LinkedIn, consulta en lote hasta 250 perfiles y exporta leads. Servidor remoto: https://mcp.linkedinenrich.com/mcp

Servidor MCP alojado

npx add-mcp 'https://mcp.linkedinenrich.com/mcp'

Se instala en Claude Code, Codex, Cursor y más

Documentación

Heads up:

The legacy automation.sayf.in/webhook/extractor endpoint is being sunset on 30 June 2026. Migrate to the v1 API on build.thepeakai.com documented below — same lookup types, new base URL + Bearer auth. /webhook/extractor on the new host is kept as a legacy alias of /api; new integrations should use /api.

Automation & API

Integrate PeakAI into your workflows with our REST API, hook into AI agents over MCP, or use our official n8n node for no-code automation.

Model Context Protocol

Plug PeakAI into Claude, Cursor & any MCP client

Our hosted MCP server exposes every contact-finding endpoint as a tool that LLM agents can call directly — no client SDK, no glue code. The agent gets your verified phone, email, and enrichment data straight from the model interface.

One URL, no install — runs entirely on our infrastructure

Authenticates with the same PeakAI token as the REST API

Streamable HTTP transport, supported by Claude, Cursor, Continue, and most agent frameworks

All v1 lookup types available as MCP tools (phone, email, enrich, DIN, reverse lookup)

Learn about MCP

MCP Server URL

Streamable HTTP transport

https://mcp.linkedinenrich.com/mcp

Paste into your MCP client's server config. Authenticate with your PeakAI access token (same one used by the REST API).

Connect your client

Tested with Claude (Desktop and Code), Cursor, and Continue. Any MCP-compatible client should work.

CClaude (Code & Desktop)

Add the server to your Claude config. In Claude Code:

claude mcp add --transport http peakai https://mcp.linkedinenrich.com/mcp \
  --header "Authorization: Bearer YOUR_ACCESS_TOKEN"

Or, for Claude Desktop, add it manually under mcpServers in claude_desktop_config.json.

⌘Cursor / Continue / others

Drop the JSON snippet into your client's MCP config:

{
  "mcpServers": {
    "peakai": {
      "url": "https://mcp.linkedinenrich.com/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_ACCESS_TOKEN"
      }
    }
  }
}

Grab your token from the REST API tab (sign-in panel) — same token works for both surfaces.

What you can ask the agent

  • • "Find the phone number for this LinkedIn profile: linkedin.com/in/kunalbahl"
  • • "Enrich these 10 LinkedIn URLs and put the results in a CSV"
  • • "Find the email for this DIN: 02890599"
  • • "Reverse-lookup [email protected] and tell me where they work"

Each call counts against your normal credit balance — the agent sees credits_remaining in every response.

Best Practices

Tips for optimal API usage

Use the Bearer header

Send the JWT as Authorization: Bearer <token> — not in the query string — so it doesn't leak into server access logs.

Set generous timeouts

At least 90 seconds. A cold provider waterfall can take 30–60s; SignalHire async is the slowest.

Refresh tokens proactively

Access tokens last 15 days. Refresh once they pass ~14 days by calling /token again (limited to 10 req/min per IP).

Cache on your side too

Our cache already deduplicates, but skipping the request entirely is faster. Cache by LinkedIn URL.

Ready to Get Started?

Contact us to enable API access on your account or discuss enterprise solutions.

Email: [email protected] • Phone: +91 9008227180