Make is the platform of choice for teams that need visual automation with more power than Zapier offers. Its scenarios handle complex branching, error handling, and data transformation beautifully. But adding AI to a Make scenario has the same problem as everywhere else: you make an HTTP call to one model at one price, with no fallback if it fails. Your beautifully designed scenario, with its error routes and conditional branches, falls over because the AI step returned a 500.
Antbase gives your Make scenarios the same resilience you build into the rest of the workflow. The HTTP module points at Antbase, and every AI call gets intelligent routing with automatic failover. A support ticket classification scenario routes to free models (classification is trivial for any LLM). A content generation scenario routes to a model that excels at writing. If any provider is down, Antbase routes around it — your scenario keeps running through its branches without hitting the error route.
For Make users specifically, the cost efficiency matters because Make charges per operation. Every failed HTTP call that triggers a retry is another operation on your plan. Antbase eliminating failures means fewer wasted operations and more predictable scenario costs. It is optimization at both the AI level and the automation platform level.
Make (formerly Integromat) is a visual automation platform that connects apps with drag-and-drop workflows. Using the HTTP module, you can call Antbase from any scenario — adding AI to your automations with intelligent model routing and no code.
HTTP Module Configuration
Add an HTTP > Make a Request module to your scenario and configure it:
- •URL: https://antbase.ai/v1/chat/completions
- •Method: POST
- •Headers: Authorization: Bearer ant_your-api-key, Content-Type: application/json
- •Body type: Raw (JSON)
{
"model": "auto",
"messages": [
{
"role": "system",
"content": "Classify the following support ticket into: billing, technical, feature-request, other"
},
{
"role": "user",
"content": "{{1.ticket_body}}"
}
],
"temperature": 0.1
}Parsing the Response
Add a JSON > Parse JSON module after the HTTP module. Map the response body to parse it. Then access the AI response via choices[0].message.content in subsequent modules.
Tips
- •Use Make variables ({{1.field}}) to inject dynamic data into your prompts.
- •For classification tasks, set temperature to 0.1 for consistent results.
- •Make has a built-in retry mechanism — set it to 1 retry since Antbase handles retries internally.
- •Create a reusable HTTP module template so you can add Antbase to any scenario quickly.
