Zhiman API platform

Build with the Micius API.

Bring Micius into your everyday R&D and operations.

Quickstart
from openai import OpenAI

client = OpenAI(
    base_url="https://api.zhiman.ai/v1",
    api_key="ZHIMAN_API_KEY",
)

response = client.chat.completions.create(
    model="micius-v1.0-flash",
    messages=[{"role": "user", "content": "Why might edge dies fail after an edge-ring change?"}],
)
print(response.choices[0].message.content)

Capabilities

Documentation

Ready for production. Integrate once and build.

  • OpenAI-compatiblePoint the OpenAI SDKs, LangChain or your own client at our base URL: change the URL, the key and the model name, nothing else.
  • Thinking controlSet a thinking budget per request, or turn thinking off for speed.
  • Tool callingParallel function calls, built for long agent loops.
  • Structured outputsResponses that always match the JSON schema you provide.
  • Confidence scoresOptional calibrated confidence for every span of the answer.
  • 256K contextWhole specifications and codebases in one request.
  • Prompt cachingRepeated context is billed at the cache-hit price.

Quick start

Your first call in a few minutes.

Pay as you go

For individual developers and small teams

  • Billed per token, top up when you need to
  • One key for the whole platform
Get an API key

Enterprise

For medium and large companies

  • Rate limits and capacity set by agreement
  • SLA, zero data retention and compliance support
Contact sales

Get your API key.

Create an account, add credits when you need them, and ship.