Make¶
A Make scenario that drafts, classifies or summarises can call your organisation's approved models through the gateway: checked against your policies, charged to the right team and recorded in Request logs, with Claude and Gemini as reachable as OpenAI's models.
What it is¶
Make (formerly Integromat) builds scenarios from modules. Its OpenAI app always calls OpenAI with an OpenAI key; its connection has no setting for another address, so it can't go through the gateway. Make's HTTP app can: its Make a request module sends a call to the gateway's chat endpoint, which reaches every provider your organisation has set up.
The two settings¶
| Setting | Value |
|---|---|
| URL | https://<your-gateway>/v1/chat/completions |
| Key | An Authorization header with the value Bearer vk_…: your Vidai key, never a provider's key |
Keep the key in a keychain rather than in the module: in Make a request, set Authentication type to API key
and, under Credentials, add a keychain that puts the key in the header, named Authorization, with the value
Bearer vk_…. Every scenario that uses the keychain then shares one place to change it.
The gateway's address is under Connect your app on the Keys page and in System › Settings › Gateway. You get the key from Keys › Issue a key; its secret is shown once.
A minimal call¶
- Add the HTTP app's Make a request module.
- Authentication type: API key, with the keychain above in Credentials.
- URL:
https://<your-gateway>/v1/chat/completions. Method:POST. -
Body content type: application/JSON. Body input method: JSON string, then the body:
{ "model": "gpt-4o-mini", "messages": [{"role": "user", "content": "Say hello in Dutch."}] } -
Parse response: Yes, so later modules can map the answer.
- Run once. The answer is in
data›choices› item 1 ›message›content.
The same module calls Claude or Gemini: put a name such as claude-sonnet-4.5 in model. The gateway translates the
request for that provider and the answer back into the OpenAI shape.
To put text from an earlier module into the body, choose Data structure as the Body input method instead of JSON string: Make then builds valid JSON even when the text has quotes or line breaks.
Reading what the gateway did¶
The module's output includes the response headers. x-vidai-request-id is the call's reference: paste it into
Request logs › Request reference. x-vidai-model says which model answered. Or find the call in Request logs
with the Key or Model filter under More filters. Cost is never in the answer; it's on the call in Request
logs.
Errors¶
With Return error if HTTP request fails on (an advanced setting), a refused call stops the module with its status, and an error handler can act on it.
| Status | Gateway code | What to do |
|---|---|---|
| 401 | invalid_api_key, missing_api_key |
Check the keychain: Bearer then the key |
| 403 | model_not_allowed, denied_by_routing_rule |
Use a model your key may call |
| 403 | content_guardrail_blocked |
A guardrail stopped the content |
| 404 | model_not_found |
Check the model name on Models |
| 429 | rate_limit_exceeded, token_limit_exceeded, concurrency_limit_exceeded |
Slow the scenario down; a Break error handler can retry later |
Route on the x-vidai-error-code header, not on the message.
Model names¶
Use the names your organisation has approved, as listed under Models in the Control room. A GET to
https://<your-gateway>/v1/models with the same keychain lists exactly the models your key may call.
What goes through the gateway¶
| In Make | Through the gateway |
|---|---|
| HTTP › Make a request calling the gateway | Yes, to every provider your organisation has set up |
| The OpenAI app and other AI apps with their own connections | No: they call their provider directly |
| Streaming | No: a module waits for the whole answer |
Check it works in a minute¶
Run the module once with the prompt "ping". A short answer means the address and key are right. The call shows in Request logs within seconds.
Good to know¶
- The gateway must be reachable from the internet: Make runs scenarios in its own cloud.
- One key per scenario, or per team of scenarios, keeps their spend apart: what the key acts for decides who pays.
- The gateway's own 429 has no
Retry-Afterheader; set a wait on your side before trying again. - A fallback chain may already have tried another model before it answers with an error; a retrying error handler on top can double the attempts.