TRAE IDE Quick Switch to Cloudwise MaaS

Add Cloudwise MaaS custom models to TRAE IDE's model manager — supporting both Anthropic-compatible and OpenAI-compatible endpoints in a few clicks

TRAE IDE allows you to configure custom third-party models alongside its built-in model library. By adding Cloudwise MaaS as a custom provider, you can use Cloudwise-hosted Anthropic, OpenAI-compatible, and other models directly inside TRAE's Builder and chat workflows.

https://www.trae.ai

TRAE — AI-Native IDE

A free AI-native IDE that automates development tasks end-to-end. Download and install TRAE before following this guide.

Prerequisites
Open Model Settings
Add an Anthropic-Compatible Model
Add an OpenAI-Compatible Model
Switch Model in Builder
Summary

Prerequisites

Requirement	Details
TRAE IDE	Installed and running on your system.
Cloudwise API Key	Create or copy your API key from Cloudwise Studio.
Anthropic Endpoint	`https://api.cloudwise.ai/api/v1/messages`
OpenAI-Compatible Endpoint	`https://api.cloudwise.ai/api/v1/chat/completions`

Open Model Settings

In TRAE IDE, open the settings panel and navigate to the Models section to manage built-in and custom models. Click the + Add Model button to start adding a new custom model.

Navigate to Settings → Models → Add Model

The model management screen lists all currently available built-in and custom models. Click + Add Model (highlighted below) to open the model configuration dialog.

Add an Anthropic-Compatible Model

Cloudwise exposes Anthropic models through a Messages API endpoint. Configure the model as shown below.

Fill in the Anthropic Model Form

In the Edit Model dialog, choose Anthropic as the provider, set the model type to Custom Model, and enter the values from the table below.

Field	Value
Provider	Anthropic
Model Type	Custom Model
Model ID	`claude-opus-4-7`
API Key	Your Cloudwise API key
Custom Request URL	`https://api.cloudwise.ai/api/v1/messages`

TRAE IDE Edit Model dialog configured for Anthropic claude-opus-4-7 via Cloudwise

Verify Supported Claude Model IDs on Cloudwise

You can look up the exact model IDs available under the Anthropic section of the Cloudwise API Reference. The screenshot below shows the supported Claude models and the confirmed POST endpoint.

Cloudwise API Reference showing Anthropic Claude model list and POST endpoint

The Cloudwise Anthropic endpoint uses the /api/v1/messages path, matching Anthropic's native Messages API. Supported models include claude-opus-4-7, claude-opus-4-6, claude-sonnet-4.6, claude-haiku-4.5, and more.

Supported Anthropic Model IDs

Model ID	Description
`claude-opus-4-7`	Latest Claude Opus — highest capability
`claude-opus-4-7-thinking`	Extended thinking variant
`claude-opus-4-6`	Claude Opus stable release
`claude-sonnet-4.6`	Fast and balanced Claude model
`claude-haiku-4.5`	Lightweight and low-latency

Add an OpenAI-Compatible Model

For OpenAI-protocol models such as GLM, GPT, and other vendors, use the OpenAI provider with the chat completions endpoint.

Fill in the OpenAI-Compatible Model Form

Click + Add Model again. In the dialog, choose OpenAI as the provider and fill in the fields below.

Field	Value
Provider	OpenAI
Model Type	Custom Model
Model ID	`glm-5.1`
API Key	Your Cloudwise API key
Custom Request URL	`https://api.cloudwise.ai/api/v1/chat/completions`

TRAE IDE Edit Model dialog configured for OpenAI glm-5.1 via Cloudwise

Verify Supported GLM / OpenAI-Compatible Model IDs on Cloudwise

Browse the GLM-5.1 (or other vendor) page in the Cloudwise API Reference to confirm available model IDs and the correct endpoint path before configuring TRAE.

Cloudwise API Reference showing GLM-5.1 model list and POST endpoint

Any model that follows the OpenAI Chat Completions protocol can use the same /api/v1/chat/completions endpoint. This includes GLM, GPT, Gemini, Grok, DeepSeek, and many others available on Cloudwise MaaS.

Important: Use the exact Custom Request URL shown above. For Anthropic models use /api/v1/messages; for all other OpenAI-compatible models use /api/v1/chat/completions. Mixing these up will result in API errors.

Suggested OpenAI-Compatible Model IDs

Model ID	Use Case
`glm-5.1`	Zhipu's latest flagship for chat and agentic tasks
`gpt-5.4`	Advanced GPT model for complex reasoning
`gpt-5`	Balanced general-purpose coding and reasoning
`gpt-4o`	Fast multimodal interactions
`deepseek-r2`	Open-source reasoning champion

Switch Model in Builder

Once your custom models are saved, you can see all of them listed under the Custom section in the TRAE model management page.

Confirm All Custom Models Are Listed

After adding both Anthropic and OpenAI-compatible models, verify they all appear in Settings → Models under the Custom section with their toggles enabled.

TRAE IDE model settings showing all added custom Cloudwise models in the list

Select Your Model in Builder Mode

Open a chat or Builder session in TRAE. Click the model selector at the bottom of the chat panel, then choose your newly added Cloudwise model from the Custom section of the dropdown.

TRAE IDE Builder mode with Cloudwise custom model selected in the model dropdown

All custom models appear under the Custom section in the model picker. Built-in TRAE models (Gemini, Kimi, Grok) remain available and unaffected by adding custom models.

Summary

You now have a complete English setup flow for TRAE IDE with Cloudwise MaaS:

Open Settings → Models in TRAE IDE
Click + Add Model and select Anthropic provider for Claude models — use endpoint https://api.cloudwise.ai/api/v1/messages
Click + Add Model again and select OpenAI provider for GLM, GPT and other models — use endpoint https://api.cloudwise.ai/api/v1/chat/completions
Enter your Cloudwise API key and the correct model ID in each dialog
Switch to a Cloudwise model in any Builder or chat session to start using it

Menu