Make.com OpenAI Integration: GPT-Powered Automation Scenarios

The Make.com OpenAI Module

Make.com has one of the best native OpenAI integrations. The module supports:

• Create a Completion: ChatGPT text generation
• Create a Chat Completion: Multi-turn conversation (better for most tasks)
• Create an Image: DALL-E image generation
• Create a Transcription: Whisper speech-to-text
• Create a Translation: Audio translation to English
• List Files, Upload File, Delete File: File management for Assistants API
• Create an Assistant / Thread / Message / Run: Full Assistants API support
• Create Embeddings: Text embeddings for semantic search

Setting Up the Connection

1. Add OpenAI module to your Scenario
2. Click "Create a connection"
3. Enter a connection name
4. Enter your OpenAI API key
5. Click Save

The connection persists. All future OpenAI modules in any Scenario reuse it.

Chat Completion — Main Action

"Create a Chat Completion" is the primary action for most AI use cases:

Configuration:

• Model: gpt-4o (most capable), gpt-4o-mini (fast/cheap), o3-mini (reasoning)
• Messages: Add rows for system prompt and user messages
• Max Tokens: limit output length
• Temperature: 0 = deterministic, 1 = creative

System + User message structure:

Output: Access response text via {{OpenAI_module.choices[].message.content}}

DALL-E Image Generation

Generate images automatically in your Scenarios:

Configuration:

• Prompt: description of the image (can use mapped data: Product: {{1.name}} in lifestyle setting, white background)
• Model: dall-e-3 (highest quality) or dall-e-2 (cheaper)
• Size: 1024×1024, 1024×1792, or 1792×1024
• Quality: standard or hd
• Style: natural or vivid

Output: URL of generated image. Next module: HTTP (GET) to download the image, or store the URL directly.

Practical workflow:

1. New product in Airtable
2. OpenAI: Create Image from product name + description
3. HTTP: Download image from URL
4. Airtable: Update record with image (upload as attachment)

Whisper Speech-to-Text

Transcribe audio automatically:

Configuration:

• File: audio file from previous module (Google Drive, Dropbox, HTTP download, etc.)
• Language: leave blank for auto-detect
• Prompt: optional context to improve accuracy (proper nouns, jargon)

Output: {{OpenAI_module.text}} — plain text transcript

Practical workflow:

1. Google Drive: new audio file in /recordings/ folder
2. OpenAI: Create Transcription
3. OpenAI (second call): "Extract action items from this meeting transcript"
4. Google Docs: Create meeting notes document
5. Email: Send summary to attendees

AI Content Pipeline with Native Module

The native OpenAI module makes building content automation fast:

Complete scenario:

1. Airtable trigger: new row with keyword + audience + tone
2. OpenAI: "Write a 1500-word SEO article targeting: {{keyword}}"
3. OpenAI (second call): "Generate meta title and description for this article"
4. WordPress: Create draft post
5. Buffer: Schedule social media posts from article content

Because the OpenAI module handles auth and response parsing automatically, this scenario has fewer steps than the equivalent with the HTTP module for Claude.

Assistants API for Complex Use Cases

For more advanced use cases — chatbots with memory, file analysis, custom tools — use the Assistants API modules:

1. Create an Assistant: Define the AI's role, instructions, and available tools
2. Create a Thread: Start a conversation
3. Create a Message: Add user message to thread
4. Create a Run: Execute the assistant
5. Get Run: Poll until complete
6. List Messages: Retrieve the response

The Assistants API maintains conversation history automatically — you don't need a Data Store for chat history.

Cost Management

OpenAI pricing varies by model:

• GPT-4o: ~$2.50/$10.00 per 1M tokens (in/out)
• GPT-4o mini: ~$0.15/$0.60 per 1M tokens
• DALL-E 3: ~$0.040-0.080 per image
• Whisper: ~$0.006 per minute of audio

For cost-sensitive workflows, use GPT-4o mini for simple tasks (classification, short responses) and GPT-4o only for quality-critical outputs. This can reduce AI costs by 5-10x.

OpenAI vs Claude in Make.com

For most Make.com users, OpenAI's native module is the easier starting point. For longer documents or better instruction following, switch to Claude via HTTP.

Recommended Tools

• Make.com — Platform with native OpenAI integration
• OpenAI API — GPT-4o, DALL-E 3, Whisper
• Claude API — Best for instruction-heavy tasks and long docs
• Airtable — Data management for AI workflows
• HubSpot — CRM for AI-powered sales automation