llm-api-adapter-zai

Official direct Z.ai / GLM Chat Completions API support for llm-api-adapter. The package targets Core >=0.9.7,<1.0.0 and does not install a provider SDK.

Installation

Install through the Core package extra:

pip install "llm-api-adapter[zai]"

Direct installation is also supported when Core is managed separately:

pip install llm-api-adapter-zai

Async requests need HTTPX:

pip install "llm-api-adapter[zai,async]"

Synchronous requests use requests by default. To opt into HTTPX for sync chat() and stream_chat(), install "llm-api-adapter[zai,httpx]" and pass transport="httpx".

Quick start

import os

from llm_api_adapter.models.messages.chat_message import UserMessage
from llm_api_adapter.universal_adapter import UniversalLLMAPIAdapter

adapter = UniversalLLMAPIAdapter(
    organization="zai",
    model="glm-5.3-flash",
    api_key=os.environ["ZAI_API_KEY"],
)

response = adapter.chat(
    messages=[UserMessage("Explain retrieval-augmented generation.")],
    max_tokens=128,
)
print(response.content)

Supported model and capabilities

The package deliberately exposes one exact, verified model ID: glm-5.3-flash.

Unknown, retired, aliased, or unverified model IDs are not inferred. Unsupported tool choices, parallel-tool control, provider-built-in tools, and every DocumentPart URL or byte form fail locally before the transport is called. The adapter accepts only the official endpoint and the ZAI_API_KEY bearer credential.

See the main llm-api-adapter README for the shared API contract and transport behavior.