Equibles serves overlapping data through two interfaces. Which you want depends on who's calling.
Pick one
| MCP server | REST API | |
|---|---|---|
| Best for | AI agents & assistants (ChatGPT, Claude, Cursor) | Your own scripts, backends, dashboards |
| You write | nothing — the model calls tools in chat | HTTP requests + JSON parsing |
| Shape | tools the model picks by name | endpoints you call by URL |
| Auth | OAuth sign-in or API key | API key |
| Endpoint | https://mcp.equibles.com/mcp |
https://api.equibles.com/v1 |
Reach for the MCP server when a person or an agent is asking questions in natural language and you want the answers in the conversation — no glue code. It's the fastest path and how most people start. See Connect the MCP server.
Reach for the REST API when you're writing code that needs the data on a schedule or inside your own product — a nightly job, a screener, a chart. Standard JSON over HTTPS. See the REST API guide.
They share one daily limit
Both surfaces draw on the same daily request quota for your account — an MCP tool call and a REST call each count once. Mixing them is fine; just budget against the single limit. See Rate limits & errors.
Using both
Nothing stops you from doing both — connect an agent over MCP for exploration, and call REST from a backend for the parts you've automated. They share the same account and key, with overlapping data; check each tool or endpoint for its exact response shape and coverage. For example, customer concentration over MCP also serves supported tagged XBRL, while REST serves verified narrative disclosures.