OpenAI-compatible API

If your tool speaks the OpenAI Chat Completions protocol, you can query your M365 data without configuring MCP at all. The serve mode runs a local HTTP server that any OpenAI-compatible client can talk to.

Start the server

npx -y @stuffbucket/workiq-proxy serve

The server starts on port 11435 by default and seeks the next available port if busy.

Custom port

npx -y @stuffbucket/workiq-proxy serve --port 8080

Endpoint

POST http://localhost:11435/v1/chat/completions

Request

{
  "model": "workiq",
  "messages": [
    { "role": "user", "content": "What emails did I get today?" }
  ]
}

Response

Standard OpenAI Chat Completions response format with the Work IQ answer in choices[0].message.content.

Use with any OpenAI-compatible client

Point any tool at http://localhost:11435 as the base URL. No API key required (the server uses your local Work IQ auth tokens).

Example: curl

curl http://localhost:11435/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"workiq","messages":[{"role":"user","content":"What meetings do I have tomorrow?"}]}'

Example: Python (openai library)

from openai import OpenAI

client = OpenAI(base_url="http://localhost:11435/v1", api_key="unused")

response = client.chat.completions.create(
    model="workiq",
    messages=[{"role": "user", "content": "Find my recent documents about budgets"}],
)
print(response.choices[0].message.content)