API Documentation

The Answira API is fully compatible with the OpenAI API format. You can use any OpenAI SDK or library by simply changing the base URL.

Introduction

Our API provides access to GLM-4.7, a state-of-the-art reasoning model with extended thinking capabilities. The API supports:

• Chat completions with streaming
• Function/tool calling
• Reasoning output (thinking process)
• JSON mode and structured outputs
• Prompt caching for reduced input costs
• Long context (up to 131K tokens)
• Request cancellation

Authentication

Authenticate requests using an API key in the Authorization header:

Authorization: Bearer YOUR_API_KEY

Sign up at the Answira Portal to get your API key.

Base URL

https://answira.ai/api

All API endpoints are relative to this base URL.

Quick Start

from openai import OpenAI

client = OpenAI(
    base_url="https://answira.ai/api/v1",
    api_key="your-api-key"
)

response = client.chat.completions.create(
    model="zai-org/GLM-4.7-FP8",
    messages=[
        {"role": "user", "content": "Hello!"}
    ],
    max_tokens=500
)

print(response.choices[0].message.content)

curl https://answira.ai/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "zai-org/GLM-4.7-FP8",
    "messages": [{"role": "user", "content": "Hello!"}],
    "max_tokens": 500
  }'

import OpenAI from 'openai';

const client = new OpenAI({
    baseURL: 'https://answira.ai/api/v1',
    apiKey: 'your-api-key'
});

const response = await client.chat.completions.create({
    model: 'zai-org/GLM-4.7-FP8',
    messages: [{role: 'user', content: 'Hello!'}],
    max_tokens: 500
});

console.log(response.choices[0].message.content);

List Models

{
  "data": [{
    "id": "zai-org/GLM-4.7-FP8",
    "name": "GLM-4.7 FP8",
    "created": 1737100800,
    "context_length": 131072,
    "max_completion_tokens": 131072,
    "input_modalities": ["text"],
    "output_modalities": ["text"],
    "quantization": "fp8",
    "pricing": {
      "prompt": "0.000000475",
      "completion": "0.000002",
      "input_cache_read": "0.00000008"
    },
    "supported_sampling_parameters": [
      "temperature", "top_p", "top_k", "frequency_penalty",
      "presence_penalty", "repetition_penalty", "stop", "seed"
    ],
    "supported_features": ["tools", "json_mode", "structured_outputs", "reasoning"],
    "datacenters": [{"country_code": "CZ"}]
  }]
}

Chat Completions

{
  "model": "zai-org/GLM-4.7-FP8",
  "messages": [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "What is the capital of France?"}
  ],
  "max_tokens": 500,
  "temperature": 0.7
}

Completions (Legacy)

{
  "model": "zai-org/GLM-4.7-FP8",
  "prompt": "Once upon a time",
  "max_tokens": 200,
  "temperature": 0.7
}

Streaming

Enable streaming by setting stream: true. The response will be sent as Server-Sent Events (SSE).

from openai import OpenAI

client = OpenAI(
    base_url="https://answira.ai/api/v1",
    api_key="your-api-key"
)

stream = client.chat.completions.create(
    model="zai-org/GLM-4.7-FP8",
    messages=[{"role": "user", "content": "Write a story"}],
    stream=True
)

for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")

Tool Calling

The model supports function/tool calling in OpenAI-compatible format.

tools = [{
    "type": "function",
    "function": {
        "name": "get_weather",
        "description": "Get current weather for a city",
        "parameters": {
            "type": "object",
            "properties": {
                "city": {"type": "string"}
            },
            "required": ["city"]
        }
    }
}]

response = client.chat.completions.create(
    model="zai-org/GLM-4.7-FP8",
    messages=[{"role": "user", "content": "What's the weather in Tokyo?"}],
    tools=tools,
    tool_choice="auto"
)

Reasoning Output

GLM-4.7 is a reasoning model that shows its thinking process. The reasoning is returned in the reasoning_content field.

JSON Mode

Set response_format: {"type": "json_object"} to constrain the model to output valid JSON.

response = client.chat.completions.create(
    model="zai-org/GLM-4.7-FP8",
    messages=[
        {"role": "system", "content": "Respond in JSON format."},
        {"role": "user", "content": "List 3 European capitals with their countries."}
    ],
    response_format={"type": "json_object"}
)

Structured Outputs

Use JSON Schema to constrain the output to a specific structure. This guarantees the response matches your schema.

response = client.chat.completions.create(
    model="zai-org/GLM-4.7-FP8",
    messages=[
        {"role": "user", "content": "Extract: John is 30 years old and lives in Prague."}
    ],
    response_format={
        "type": "json_schema",
        "json_schema": {
            "name": "person_info",
            "schema": {
                "type": "object",
                "properties": {
                    "name": {"type": "string"},
                    "age": {"type": "integer"},
                    "city": {"type": "string"}
                },
                "required": ["name", "age", "city"]
            }
        }
    }
)

Prompt Caching

Repeated prompt prefixes may be served from cache at a reduced price. Cached input tokens cost $0.08/M compared to $0.475/M for regular input tokens.

The usage object in the response includes prompt_tokens_details.cached_tokens showing how many tokens were served from cache:

// In the response usage object:
{
  "usage": {
    "prompt_tokens": 1500,
    "completion_tokens": 200,
    "total_tokens": 1700,
    "prompt_tokens_details": {
      "cached_tokens": 1200
    }
  }
}

Request Cancellation

You can abort a running request at any time by closing the HTTP connection. The server detects the disconnection and immediately stops generation on the backend, freeing GPU resources.

You are only billed for tokens actually generated before the cancellation.

Error Handling

The API returns standard HTTP error codes:

Rate Limits

Rate limits are applied based on server capacity. When the server is at high load, requests may receive a 429 response with a Retry-After header.