Platform Integration (server-side)
This page is for fintech platforms, brokers, robo-advisers, competition and education products, and data apps that want RiskModels numbers inside their own product. It covers the setup where your servers call RiskModels and your users never see a RiskModels sign-in.
For AI assistants used by one person (Claude, ChatGPT, Cursor, Gemini) see Agent integration; the MCP server there is built for one person signing in with their own account.
How it works
- One organization key. Create one API key for your organization and keep it on your servers, for example in the service that already holds your other market-data vendor keys. Your end users do not need RiskModels accounts. The key must never reach a browser, a mobile app or any client-side code.
- Call the REST API from your backend. Every MCP tool has a REST equivalent (API reference, OpenAPI). Typical calls for a platform:
POST /api/portfolio/exposure: a book's market, sector, subsector and stock-specific exposure, long or short, up to 1,000 names.POST /api/decompose: one stock's risk split with ETF hedge ratios.GET /api/metrics/{ticker}: latest metrics for one stock.GET /api/identity/resolve: which security a ticker meant on a date and what a holding became (free).
- Cache, then serve your users from your cache. The data is end-of-day, so one response per security (or per book) per trading day is enough. Refresh after each US market close.
- Show the attribution. Wherever a screen, statement or PDF shows RiskModels numbers, show "Powered by RiskModels" or "Risk data provided by RiskModels".
What the terms let you cache
Under the API terms, the licence lets you integrate the API into your applications and AI agents and use its output in your analyses, reports and products. In practice:
| You may | You may not, without a separate written agreement |
|---|---|
| Store the responses you retrieve, keyed however suits you, for example by security and date | Offer RiskModels numbers to third parties as a feed, database, bulk download or API, including a public or partner API you run |
| Serve cached numbers to your own users inside your product: screens, alerts, statements, PDFs, your own AI assistant's answers | Resell RiskModels output as a standalone data product |
| Keep the history you have fetched for as long as your licence is active | Train models on RiskModels output to replicate or compete with our risk models |
| Use one organization key on your servers for all your users | Give the key to users or embed it in client code |
When you stop using the API, delete the cached output (API terms, section 11).
Redistribution licence for platforms
If you want to carry RiskModels numbers through your own data API, signals product or partner feed, we offer a redistribution licence for platforms, with attribution. Contact service@riskmodels.app with the surfaces you have in mind.
Cost
Pricing is per call (see pricing), so caching sets your cost. Fetching only the stocks and books your users actually hold, once a day, can cost a few dollars a day. Re-fetching the whole universe every day costs much more and is rarely needed.
Checklist
- One organization key, held server-side, rotated if exposed.
- Results cached per security or book per trading day, refreshed after the US close.
- "Powered by RiskModels" shown wherever the numbers appear.
- No RiskModels numbers in feeds or APIs you offer to others unless you have a redistribution licence.
- US equities and ETFs only; end-of-day, not intraday.