Integrations
REST, MCP,
webhooks and CSV
ProspectAPIs is a prospect data API you can reach four ways: the REST API for your own code, with JavaScript and Python clients; the MCP server, so an AI agent can call every endpoint as a tool; signed webhooks that push new funding rounds to you; and CSV for spreadsheets and list cleaning. One key and one prepaid balance work across all four, and no method costs extra.
Four ways to receive the data
Pick the one that fits the job. Every method bills like the REST call it makes.
| Method | How it works | Best for | Freshness | Start with |
|---|---|---|---|---|
| REST API | You pull. Send a request with your key and get JSON back. | Enrichment by domain, scheduled jobs and your own product. | Live at the moment you call. | https://api.prospectapis.com/v1 with Authorization: Bearer |
| MCP server | Your agent pulls. Each endpoint is a tool it can call. | AI agents and assistants doing research or outreach. | Live at the moment the tool runs. | npx -y prospectapis-mcp, or the hosted server with OAuth |
| Webhooks | We push. A watchlist posts each new matching funding round to your URL, signed. | Alerts, CRM sync and Slack or no-code triggers. | On the delivery pass after a round lands. | POST /v1/watchlists with filters and a webhook_url |
| CSV | You download. Email verification jobs return their results as CSV, and a sample script writes funding rounds to CSV. | Spreadsheets, list cleaning and one-off exports. | As of the job or the run. | format=csv on a verification job's results |
The numbers
Per our API reference, every key gets 60 requests a minute on REST and a separate 60 on MCP, and every method draws from the same balance.
- MCP tools
- 99
- One per endpoint, with its price in the tool text.
- Requests per key
- 60 / min
- Separate counters for REST and MCP.
- Client libraries
- JS, Python
- JavaScript on npm, and a one-file Python client.
- Free to start
- $1
- No card, and it works on every method.
REST API and client libraries
Base URL https://api.prospectapis.com/v1, a bearer key on every request, JSON in and out. Install the JavaScript client with npm install prospectapis, or download the Python client as one file. The OpenAPI document imports into Postman or Insomnia and generates a typed client in any language.
curl "https://api.prospectapis.com/v1/funding?round=series_a&limit=5" \ -H "Authorization: Bearer $PROSPECTAPIS_KEY"
| Route | What it does | Price |
|---|---|---|
| GET /v1/funding | Funding rounds, filtered and paged. | Loading the live price |
| POST /v1/research | A cited brief on one company or person. | Loading the live price |
| POST /v1/email/verify | Check 1 or more addresses before you send. | Loading the live price |
| POST /v1/watchlists | Have new funding rounds pushed to your webhook. | Loading the live price |
The routes most teams start with, priced from our live price list; the API reference lists every route with its cost.
MCP server for AI agents
The prospectapis-mcp server turns every endpoint into a tool. Add it to Claude Code with one command:
claude mcp add prospectapis -e PROSPECTAPIS_API_KEY=pa_live_... -- npx -y prospectapis-mcp
- Any local MCP client. Claude Desktop, Cursor, Codex and others run
npx -y prospectapis-mcpwith your key in the environment. - Remote clients with OAuth. Claude.ai, ChatGPT and other clients that support remote MCP servers add the hosted server and sign in instead of pasting a key.
- Same prices. Each tool bills exactly like the API call it makes, and the balance and price tools are free.
Setup for each client is on the MCP server page, and our guide to researching prospects with an AI agent shows an agent using it.
Webhooks
A funding watchlist posts each new matching round to an https URL you choose, signed with HMAC-SHA256 and billed only when your server answers 2xx. The URL can be your own endpoint or a webhook trigger in a no-code tool.
Zapier, Make and n8n
There is no native app to install. Use the tool's HTTP step to call any route, and its webhook trigger to receive watchlist deliveries:
- Call the API. An HTTP request step with
Authorization: Bearerand your key, then map fields from the JSON answer. - Receive new rounds. Create a watchlist whose webhook_url is the tool's webhook trigger, then send each round to Slack, a sheet or your CRM.
- Verify emails in a flow. Send each new lead to
POST /v1/email/verifyand branch on its status and recommendation.
CSV
An email verification job returns its results as CSV with format=csv. For funding rounds, this Python script pages through a search and writes a CSV, stopping cleanly if the balance runs out:
import csv, datetime, os, requests
since = (datetime.date.today() - datetime.timedelta(days=30)).isoformat()
headers = {"Authorization": f"Bearer {os.environ['PROSPECTAPIS_KEY']}"}
params = {"round": "series_a", "since": since, "limit": 100}
rows = []
MAX_PAGES = 10 # up to 1,000 rounds; raise it on purpose, each record is billed
for _ in range(MAX_PAGES):
res = requests.get("https://api.prospectapis.com/v1/funding", headers=headers, params=params)
if res.status_code == 402:
print(f"out of credits, saved {len(rows)} rows")
break
res.raise_for_status()
page = res.json()
rows += page["data"]
if not page.get("next_cursor"):
break
params["cursor"] = page["next_cursor"]
with open("series_a.csv", "w", newline="") as f:
w = csv.writer(f)
w.writerow(["company", "domain", "amount_usd", "announced_at", "investors", "source"])
for r in rows:
w.writerow([r["company_name"], r["company_domain"], r["amount_usd"], r["announced_at"],
"; ".join(r.get("investors") or []), r["source_url"]])
print(f"wrote {len(rows)} rounds to series_a.csv")Questions
- Do you have a Zapier, Make or n8n app?
- Not a native app. Each of those tools has an HTTP request step and a webhook trigger, and both work with the API today: send requests with your key in the Authorization header, and point a funding watchlist at the webhook trigger.
- Is there an SDK?
- Yes. A JavaScript client on npm as prospectapis, and a one-file Python client you can download from this site. Both wrap the same REST routes, and the OpenAPI document lets you generate a client in any other language.
- Which AI tools can use the MCP server?
- Any MCP client. Run the server locally with npx in Claude Code, Claude Desktop, Cursor or Codex, or add the hosted server to a client that supports remote MCP servers, such as Claude.ai or ChatGPT, and sign in with OAuth.
- Can I get the data as a spreadsheet?
- Email verification jobs return their results as CSV. For funding rounds, our sample Python script pages through a search and writes a CSV, so you control exactly which rounds you pay for.
- Do integrations cost extra?
- No. Every method bills exactly like the REST call underneath it, from the same prepaid balance. Installing the MCP server, creating a watchlist and downloading the SDKs are free.
More answers on the FAQ page.