Reads stay bounded
Autonomy policy: account reads and cached research stay inside the written pilot scope.
leadhoundapi
Choose a plan
Connect leadhoundapi (leadhoundapi.com) to your agent so it can read local markets, rank recent public service-gap reports, and draft human-reviewed outreach openers. Signals are not guarantees or proof of intent.
Works with:
Plans from $39/mo · checkout is live
npm install -g leadhoundapi leadhoundapi auth login lh_live_... leadhoundapi mcp install claude --run
Hosted MCP clients can connect to https://api.leadhoundapi.com/mcp with OAuth or an account key.
Autonomy policy: account reads and cached research stay inside the written pilot scope.
Paid pilot calls and jobs have explicit limits. No tool can create unlimited vendor cost or a surprise overage.
Keep a human on send. leadhoundapi gives evidence for outreach, not proof of intent.
claude mcp add leadhoundapi -t http https://api.leadhoundapi.com/mcp -H "Authorization: Bearer lh_live_YOUR_KEY"
Settings → Connectors → Add custom connector → paste https://api.leadhoundapi.com/mcp. OAuth signs you in; no key handling.
Settings → Connectors → Add → https://api.leadhoundapi.com/mcp. Works in chat and deep research modes via OAuth.
{ "mcpServers": { "leadhoundapi": {
"url": "https://api.leadhoundapi.com/mcp",
"headers": { "Authorization": "Bearer lh_live_YOUR_KEY" } } } }Apps and agents built in Replit or Lovable speak MCP with the config above, or call REST directly; the API quickstarts have a paste-ready fetch snippet.
leadhoundapi find "roofer" "Dallas, TX" --json
Full command set on the Agent CLI page.
The same agent that built your workflow can research public service-gap reports. Ask for a city, vertical, pattern, and opener. The agent stays inside the account's pilot scope.
User: Find appliance repair shops in Phoenix whose recent public reviews report missed calls. Agent: I will call leadhoundapi and cap spend to one paid market read. Tool: market scan returned recent public review patterns and provenance. Agent: Top research hypothesis is missed-call recovery plus warranty follow-up. I will draft one opener for human review. User: Export JSON and wait before sending. Agent: Export ready. I will not send outreach without approval.