The whole platform,
as an endpoint.
A REST API over every metric and all of the history, an MCP server your assistant can call directly, and an alerting system that will post to your own endpoint. Every request, tool call and delivery is logged with its response.
API and MCP are on Vega · Alerts start on Gamma
$ curl https://optionomics.ai/api/v1/overview \-H "X-USER-EMAIL: [email protected]" \-H "X-USER-TOKEN: $OPTIONOMICS_API_KEY"{"session": "open","regime": "risk_on","net_premium": 418204119,"put_call_ratio": 0.82}
Give your agent
somewhere to look.
Point Claude, Cursor or ChatGPT at one URL and your assistant can read the options market for itself — the same readings the screens draw, reached the way it already reaches everything else.
One config block, then ask in English.
"What does the unusual activity in NVDA look like this week, and how has that pattern resolved before?" Your assistant works out what it needs and goes and gets it — the scored tape, then the flow behind it, then the record of how that shape resolved before. You never open a tab.
The server speaks streamable HTTP, and what it offers is stable, so a client that caches it does not go stale.
{"mcpServers": {"optionomics": {"type": "http","url": "https://optionomics.ai/mcp","headers": {"X-USER-EMAIL": "[email protected]","X-USER-TOKEN": "$OPTIONOMICS_API_KEY"}}}}
Things people ask it:
- What unusual activity is there in NVDA this week? Reads the scored tape, not the raw prints
- Has this pattern resolved before, and how? Goes to the public record and comes back with it
- Where is the gamma sitting on SPY right now? Dealer exposure across strikes, this session
- Which of my watchlist names had congressional filings? Form 4 and PTR filings, matched to the symbol
- Summarise the flow in semis since the open. Net premium, direction and where it turned
Skip the OPRA pipeline.
The same data the screens read, over HTTP, with a published OpenAPI specification and a reference you can send requests from.
- Market data
- Overview, tickers, chains, derived metrics, flow aggregates, unusual activity and levels — the readings, not the raw feed you would have to build them from.
- Your own resources
- Alerts, notifications, push devices, chats and subscription status.
- Keys you control
- Issue and revoke API keys from the developer console. Every REST call is logged as its own row, with the arguments and the response.
- A specification, not a wiki
- The OpenAPI document is served from the API itself, so a generated client is generated from the thing that is actually running.
Say it once.
It watches from then on.
The live streams, your conditions, the channels you chose — and a record of every attempt, including what your endpoint said back.
- Watch what matters
- Unusual option prints, trade ideas, market commentary, flow acceleration, and insider or congressional filings. Templates get you started; the conditions are yours.
- Webhooks with receipts
- Post to your own endpoint and read back every delivery: the payload, the status code, and the body your server returned. A silent integration is a thing you can look at rather than a thing you suspect.
- One event, one message
- A print that raises several of your rules notifies once, and a batch of filings arrives as one message rather than forty. That is the database's job, not a retry loop's.
- Cooldowns that behave
- A cooldown counts from the last thing that fired, not the last thing that matched — so a steady stream of matches cannot extend the silence forever.
See what your code did.
Keys, the REST request log, the MCP tool-call log, the connected clients and the reference — five pages behind one header, inside your account.
Get a key.
Vega opens the REST API, the MCP server and backtesting on top of everything else. Alerts and webhooks start one rung lower, on Gamma.
Cancel anytime · Month to month · Switch tiers whenever