Quickstart
Connect Edgrapi to your AI agent, or call the REST API directly — a few minutes either way.
1. Get an API key
Create an account with an email and password, no card — a key is minted for you
on signup, and you get 100 free credits every month. Make more keys under Dashboard → API keys.
Keys look like edgr_AbCd… — use it as a Bearer token for MCP, or in the X-API-Key header for REST.
export EDGRAPI_KEY=edgr_...
2. Connect your agent (MCP)
The fastest path to value: point Claude, Cursor, Cline or your own agent at the hosted MCP server. It gets the
full toolset — get_insider, get_holdings (13F), get_activist,
get_events (8-K), get_fundamentals and more — and calls Edgrapi on every question,
which is the loop that keeps working long after a one-off script would be done.
{
"mcpServers": {
"edgrapi": {
"url": "https://api.edgrapi.com/mcp",
"headers": { "Authorization": "Bearer {EDGRAPI_KEY}" }
}
}
}
Then just ask, in plain English: "What did NVIDIA insiders buy or sell recently, and how fast is revenue
growing?" — the agent calls get_insider and get_ratios and answers. Full setup
per client (OAuth 2.1 or Bearer) is in AI agents & MCP.
3. Prefer REST? Make a call
Every tool is also a plain REST endpoint. Pull NVIDIA's parsed insider trades — the smart-money layer a free SEC library won't give you. Each Form 4 comes back with the transaction code decoded, a buy/sell signal, and the dollar value computed:
curl "https://api.edgrapi.com/v1/insider/NVDA" \
-H "X-API-Key: $EDGRAPI_KEY"
import os, requests
r = requests.get("https://api.edgrapi.com/v1/insider/NVDA",
headers={"X-API-Key": os.environ["EDGRAPI_KEY"]}, timeout=60)
r.raise_for_status()
t = r.json()["filings"][0]["transactions"][0]
print(t["signal"], t["shares"], t["value"])
const res = await fetch("https://api.edgrapi.com/v1/insider/NVDA",
{ headers: { "X-API-Key": process.env.EDGRAPI_KEY } });
const data = await res.json();
console.log(data.filings[0].transactions[0].signal);
A successful response:
{
"ticker": "NVDA", "company": "NVIDIA CORP", "count": 1,
"filings": [
{ "form": "4", "filed": "2026-09-02",
"owner": "STEVENS MARK A", "relationship": ["director"],
"transactions": [
{ "code": "S", "code_label": "Open-market sale", "signal": "sell",
"shares": 447400, "price_per_share": 220.10, "value": 98470905.66 }
] }
],
"source": "SEC EDGAR Form 4"
}
4. Company financials
Fundamentals and ratios are here too — flat, computed, ready to drop into a DataFrame:
# Income statement, balance sheet, cash flow as flat named fields
curl "https://api.edgrapi.com/v1/fundamentals/AAPL?period=annual&limit=4" -H "X-API-Key: $EDGRAPI_KEY"
# Margins, ROE/ROA, leverage and liquidity, already computed
curl https://api.edgrapi.com/v1/ratios/AAPL -H "X-API-Key: $EDGRAPI_KEY"
5. List recent filings
curl "https://api.edgrapi.com/v1/filings/AAPL?form=10-K" -H "X-API-Key: $EDGRAPI_KEY"
Each plan meters total requests; free accounts get 100 to start. See Rate limits & plans.
6. More of the smart-money layer
Your first call already pulled parsed insider trades. The rest of the smart-money layer works the same way — institutional holdings, activist stakes and 8-K events, normalized so you don't build the parser:
# A fund's latest 13F portfolio, diffed against last quarter
curl "https://api.edgrapi.com/v1/holdings/berkshire" -H "X-API-Key: $EDGRAPI_KEY"
# 13D/13G filings — anyone crossing 5% of a company
curl "https://api.edgrapi.com/v1/activist/latest?activist_only=true" -H "X-API-Key: $EDGRAPI_KEY"
# 8-K material events, tagged by item code
curl "https://api.edgrapi.com/v1/events/AAPL?notable=true" -H "X-API-Key: $EDGRAPI_KEY"
Next
- Insider trades, 13F holdings, activist stakes: the smart-money endpoints in full
- Authentication: keys, RapidAPI, and MCP bearer tokens
- AI agents & MCP: using Edgrapi from Claude, Cursor and other MCP clients
- Full API reference: every endpoint and field