API
Your workflows, callable over HTTP.
Switch a workflow to App mode and it becomes an HTTP route. Your backend, n8n or Make runs it with a key, sets a credit cap and gets back the images or videos it produced.
- No subscription for personal accounts
- Same price as in the app
- 60 requests per minute per key
curl -X POST "$BASE/v1/apps/proj_3d90b7:wf_8f2c1a4e/runs" \
-H "Authorization: Bearer $BEEMM_KEY" \
-H "Idempotency-Key: order-4821" \
-H "Content-Type: application/json" \
-d '{"inputs":{"prompt":"camel leather bag, studio background"},"maxCredits":20}'202 response
{ "id": "api7c1e…", "object": "run", "status": "queued", "pollAfterSeconds": 3 }From the editor to production in three steps
- 01Turn on App modeIn the workflow editor. Give the Input role to the nodes that become parameters (text, number, upload…), and the Output role to the ones whose result you want. Nothing to publish.
- 02Create a keySettings → Developers → API keys. Read or Execute rights, one key per integration, and a monthly allowance when the key goes to a third party.
- 03Run it, then read the resultPOST /runs returns a run ID. GET /runs/{runId} returns the status, then the URLs of the images and videos produced.
Built never to charge you twice
An integration that loops, or retries after a network drop, must not drain a wallet. These safeguards are required, not optional.
A cap on every run
maxCredits is required on every run. The run stops before the charge that would exceed it; whatever is not executed is not billed.
Retries without double billing
Same Idempotency-Key and same body: you get the existing run back, nothing is rerun or billed again.
A monthly allowance per key
A key handed to n8n or a contractor never goes past the budget you set. Change the cap without replacing the key.
A contract you can check
Every app sheet and every run carries a contractHash. It changes only when the expected parameters change, never when you move a node or switch models.
Six routes, that’s all
Bearer header authentication, errors typed by a stable code.
- GET
/v1/modelsRunnable models - GET
/v1/balanceYour balance and the key’s allowance - GET
/v1/appsYour workflows in App mode - GET
/v1/apps/{appId}An app’s parameters and outputs - POST
/v1/apps/{appId}/runsStarting a run (spends credits) - GET
/v1/runs/{runId}Status and result of a run
Frequently asked questions
- Do I need a subscription to use the Beemm Vision API?
- Not for a personal account: creating a key and running an app only need credits. Organisations keep their plan. A key with no credits can still read your apps and results, but cannot start a run.
- Can I generate an image in a single request?
- Not directly: the API runs workflows. Put the model in a workflow, even a single-node one, switch it to App mode, then run it. You choose the model, its settings and the outputs once and for all.
- How much does an API run cost?
- The same credit price as in the app, charged node by node. You set the maximum on every call with maxCredits: the run stops before going over it.
- How is it different from the MCP connector?
- The MCP connector lets an assistant such as Claude or ChatGPT drive Beemm Vision in a conversation. The API is for a program: a backend, an n8n or Make scenario, that runs your workflows unattended.
Connect your first workflow
Create a key in Settings, copy the app ID shown in the editor, run it. For high volume or an organisation, the team will get back to you.