About this project
zai-all-in-one reverse-engineers the GLM model APIs provided by Zhipu AI (智谱) and wraps them behind a single, clean endpoint. Instead of dealing with Zhipu's proprietary authentication flow and request format directly, clients talk to zai using a standard interface.
The project focuses on intercepting and replicating the internal request contracts Zhipu uses for its GLM family — GLM-4, GLM-3-Turbo, and related variants — making them accessible without the usual friction of their SDK or token management.
Key features
Reverse-engineered GLM access
Intercepts Zhipu AI's internal API contracts for GLM-4 and GLM-3-Turbo.
Single unified endpoint
All GLM variants accessible through one consistent interface.
No official SDK required
Bypasses Zhipu's proprietary auth flow and token management.
Lightweight proxy layer
Minimal setup — drop in and start querying GLM models immediately.