Using BeatAPI in Cursor

Cursor is an AI code editor. Its model settings accept an OpenAI-compatible endpoint, so BeatAPI can be supplied as a custom provider without any extension.

Prerequisites

1. Cursor installed.
2. A BeatAPI API key — create one in Dashboard → API Keys.

Step 1: Open model settings

Open Settings — the gear icon at the top right, or Cmd+Shift+J on macOS, Ctrl+Shift+J on Windows and Linux — then choose Models in the sidebar.

Step 2: Point it at BeatAPI

Find the OpenAI API Key section and enable Override OpenAI Base URL, then enter:

Save. Cursor verifies the endpoint before accepting it.

Step 3: Add the models you want

Cursor keeps its own model list, and an entry in that list is not proof the account serves it. Add the model IDs below by name, and enable only those — a model that BeatAPI does not serve returns not_found on first use.

Cursor lets you pick a different model per feature — Chat, inline edit (Cmd+K), and completion. Pairing a cheap model with completion and a strong one with Chat is the single biggest lever on cost, because completion fires far more often than chat does.

Step 4: Use it

• Chat — Cmd+L / Ctrl+L, then ask a question or describe a task.
• Inline edit — select code, press Cmd+K / Ctrl+K, describe the change.
• Completion — type; press Tab to accept a suggestion.

Reference files with @filename and directories with @folder so the model sees the right context instead of guessing from the open buffer.

FAQ

Cursor cannot reach BeatAPI

1. Base URL must be exactly https://api.beatapi.io/v1.
2. The key must be active — check Dashboard → API Keys.
3. Behind a corporate proxy, allow api.beatapi.io.

A model in the list does not work

Cursor’s list is its own. Only the IDs the account actually serves will answer; see the pricing page for the current catalogue.

402 insufficient_credits

The balance is exhausted. Add credits from the dashboard.

429 rate_limit_exceeded

The key exceeded its request rate. Completion is the usual cause, because it fires on almost every keystroke pause. Move completion to a cheaper model, or raise the allowance — the per-minute rate rises with lifetime top-ups and is shown on the dashboard.

Responses are slow

Use a lighter model for completion and inline edits, keep selections tight, and prefer @file references over pasting large blocks.

Support

Open a support ticket from the dashboard and include the request_id from the failing response.