Test Custom ChatGLM2 Models on FastGPT

API Request Overview This section details the official API request format for testing a custom ChatGLM2 model integrated with FastGPT. All requests use a POST method targeting the /v1/chat/completions endpoint on your deployed FastGPT domain.

Required Request Headers Two standard headers are mandatory for all test requests:

1. Authorization: Bearer [REDACTED_CREDENTIAL]: Preserve the sk- prefix and use either the example token or your valid custom API token for authentication.
2. Content-Type: application/json: Ensures the request body is properly parsed by the endpoint.

Step-by-Step Test Command Use the following exact curl command to send a test chat completion request. Replace https://domain with your FastGPT deployment’s public domain:

curl --location --request POST 'https://domain/v1/chat/completions' \
--header 'Authorization: Bearer [REDACTED_CREDENTIAL]' \
--header 'Content-Type: application/json' \
--data-raw '{
  "model": "chatglm2",
  "messages": [{"role": "user", "content": "Hello!"}]
}'

After executing this command, the endpoint will return a standard chat completion response compatible with FastGPT’s custom model integration.

Critical Configuration Rules There are two non-negotiable requirements for successful testing:

1. The Authorization header must use a valid token with the sk- prefix format, matching your FastGPT API credentials.
2. The model field in the request body must exactly match the custom model name configured in the One API panel. Using an incorrect model name will cause the request to fail.

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.