Integration · Developer tools
Connect Claude & ChatGPT to Azure OpenAI
Give your AI assistant access to a model you run in your own Azure OpenAI resource. It can send a conversation to your chat deployment and get the reply, turn text into embeddings with an embedding deployment, or generate images with a DALL-E deployment. Requests go to your resource, under your Azure settings.
Free plan, no credit card. Takes about a minute.
Try asking
“Ask our Azure deployment to summarize this contract in five bullet points.”
The AI sends the text to your chat deployment and returns the model's reply.
“Get a second answer to this question from our own model and compare.”
It sends the question as a conversation to your deployment and puts the reply next to its own.
“Run this prompt with temperature 0 and at most 200 tokens.”
It calls the chat deployment with those settings and returns the result.
What your AI can do
Chat completions
Send a conversation with system, user and assistant messages to your chat deployment, with optional temperature and max tokens.
Embeddings
Turn one text or a list of texts into vector embeddings with an embedding deployment.
Image generation
Create images from a text prompt with a DALL-E deployment, with size, quality and style options.
“Make embeddings for these ten product descriptions.”
It sends the texts to your embedding deployment and returns one vector per text.
“Generate a 1024x1024 image of a flat illustration of a warehouse.”
It sends the prompt to your DALL-E deployment and returns the generated image.
“Draft a system prompt, then test it against our deployment.”
It writes the prompt, sends a test conversation with a system and a user message, and shows you the reply.
3 tools for Azure OpenAI
These tools are switched on when you connect. You can switch any of them off, or require your approval before it runs.
create_chat_completionChanges dataSend a conversation to the connected chat model deployment and get its reply. Pass `messages` (list of {role, content}; roles system, user, assistant), set `api_version` to 2024-10-21, and optionally `temperature` or `max_tokens`.create_embeddingChanges dataTurn text into a vector embedding with the connected embedding deployment. Pass `input` (text or list of texts) and set `api_version` to 2024-10-21. Only works if the deployment is an embedding model.generate_imageChanges dataGenerate an image from a text `prompt` with the connected DALL-E deployment; optional `size`, `quality`, `style` and `n`. Set `api_version` to 2024-10-21. Only works if the deployment is an image model.
Set up in three steps
- 1
Pick the app
Create a free PipMCP account and choose Azure OpenAI from the app list.
- 2
Paste your key
Enter your Azure OpenAI resource name (the NAME in https://NAME.openai.azure.com), paste Key 1 or Key 2 and enter the deployment name. In the Azure portal, open your Azure OpenAI resource and go to Resource Management > Keys and Endpoint to copy the key. The deployment name is listed under Deployments in Azure AI Foundry.
- 3
Add the link to your AI
You get a personal MCP link. Add it to Claude, ChatGPT or Cursor:
- Click your name, then Settings › Connectors › Add custom connector.
- Paste your link as the Remote MCP server URL.
- Switch it on from the + menu in a chat.
Questions
What can the AI do in Azure OpenAI?
It can send conversations to your chat deployment and get replies, create embeddings with an embedding deployment and generate images with a DALL-E deployment. It works with the deployment you connect.
Does the AI see my Azure OpenAI key?
No. You paste your Azure OpenAI key into PipMCP once. It is encrypted at rest, never shown to the AI and never shown again after you save it. The AI only gets the results of the tools you have switched on.
Can I control what the AI is allowed to do?
Yes. You choose which tools are switched on, for example chat completions only. Any tool can be set to require your approval before it runs, and every tool call is logged.
Does it work with ChatGPT?
Yes. In ChatGPT, open Settings › Apps & Connectors › Advanced, turn on Developer mode and add your PipMCP link. Developer mode needs a paid ChatGPT plan: Plus, Pro, Business or Enterprise. The same link also works in Claude (Settings › Connectors › Add custom connector), Cursor and other MCP clients.
What does it cost?
PipMCP has a free plan, and you do not need a credit card to start. Paid plans are billed per completed task. The requests run on your own Azure OpenAI resource, so anything Azure charges for them is between you and Microsoft.
Why do I need a deployment name?
In Azure OpenAI you call a model through a deployment you have created in your resource. The connection points at one deployment, so the tools that work depend on the model deployed there: a chat model for chat completions, an embedding model for embeddings and a DALL-E model for images.
Related integrations
Let your AI work in Azure OpenAI today.
Start free. Your key stays encrypted, and you decide what the AI may do.
Connect Azure OpenAI freePipMCP is not affiliated with Azure OpenAI. Product names are trademarks of their owners.