Prerequisite Navigation and Model Prep
First, navigate to the Model Channels tab (shown in screenshot aiproxy-1). Only models pre-configured in the Model Configuration menu can be added to a channel; the FastGPT system includes mainstream models by default. If a required model is not available in the default list, add it via Model Configuration first before proceeding.
Step-by-Step Channel Creation
- Click the "Add Channel" button in the top-right corner of the Model Channels tab to open the configuration page (screenshot
image-122). - Use Alibaba Bailian models as a reference for filling out the configuration fields (screenshot
image-123). The required configuration fields are detailed below:
| Configuration Field | Description |
|---|---|
| Channel Name | A display-only label for channel identification |
| Protocol Type | The API protocol for the model. Most providers support the OpenAI protocol, which can be selected as a universal option |
| Models | Pre-configured models available for the channel. If the desired model is not in the dropdown menu, click "Add Model" to add a custom model via the linked configuration flow |
| Model Mapping | A JSON object mapping internal FastGPT model names to the actual model names used by the upstream provider. Example: {"gpt-4o-test": "gpt-4o"} — FastGPT uses the internal name, while AI Proxy forwards the upstream model name |
| Proxy URL | The base URL for the model API; do not input the full request URL. Confirm whether the path /v1 must be appended to the base URL |
| API Key | API credentials obtained from the model provider. Some providers require multiple keys; follow on-screen prompts for input requirements |
- Click "Add" to save the new channel, which will appear in the Model Channels list (screenshot
aiproxy-4).
Key Configuration Notes
- The Proxy URL field only accepts the base API endpoint, not the full model request path. Always verify if the
/v1path segment is required for the target provider. - Model mapping creates a consistent internal naming layer, eliminating the need to update internal FastGPT references when upstream model names change.
- All API keys must be obtained directly from the respective model provider, with credentials matching the protocol and proxy URL configured for the channel.
Source: FastGPT official source
Applicability and version scope
Use this page for the documented Deployment and upgrades scenario. Confirm the FastGPT, dependency, API, and deployment versions in the official source before applying a change.
Safety guardrails
Use [REDACTED_CREDENTIAL] for credentials and private data. Confirm the documented environment and version before review.
Rollback guidance
Restore the prior technical-content authority snapshot. Restore saved configuration and data snapshots, then repeat the smallest verification scenario.