Skip to main content
If you build with Cursor, Claude, or ChatGPT, you can use MCP for interactive prospecting and the REST API for scheduled pipelines. This page ties both surfaces together without mixing reference docs.

​
Pick your surface

You want to…Use
”Find VP Sales at fintech companies and draft email”MCP — search_contacts, research_contacts
Nightly enrich 500 accounts from a CSVREST — Enrich a lead list
Product UI with instant enrichmentREST + Webhooks
Repeatable agent with guardrailsDeterministic agent workflow

​
Example: agent prompt (MCP)

Paste into Claude Code or Cursor after connecting MCP: 
Search for CTOs at B2B SaaS companies with 51-200 employees.
Pick the best match, research the contact, and summarize:
- Name, title, company, email, phone
- One personalized opener sentence for cold email
Use Seamless MCP tools only. Stop if research fails and report credits/status.
See Prospect to meeting for a full MCP workflow with diagrams.

​
Example: automation script (REST)

For production automation, use the REST API so runs are logged, idempotent, and testable: 
# Same flow as enrich_list.py — see use-cases/enrich-a-lead-list
export SEAMLESS_API_KEY=your_key
python enrich_list.py
Follow Build a deterministic agent workflow to define state (searchResultId, requestId), stop conditions, and error handling for autonomous jobs.

​
Learn the product (non-developers)