What the data for this category looks like
Daily documents for this category include service request forms, incident tickets, processing records, and user satisfaction surveys. Data sources include ITSM systems, phone or chat logs, asset systems, and knowledge bases. Data update frequency is daily or real-time. It includes structured ticket fields such as ticket number, problem category, processing status, semi-structured chat logs, and unstructured knowledge base documents. Data fields include user identifiers, problem descriptions, processing progress, satisfaction scores, and other relevant items.
What constraints these characteristics impose on the model access and configuration link
Data sources for this category are diverse, including structured tickets, semi-structured chat logs, and unstructured knowledge base documents. The model access link must support multi-format data processing. Real-time updated data requires configuration that supports real-time context synchronization to match the real-time requirements of ticket processing. Ticket fields include ticket number, problem description, processing status, user satisfaction scores, and other relevant items. The model must accurately map fields to corresponding categories. If an error occurs, employees may be unable to perform their work, or SLA deadlines may be missed. Therefore, stability configuration for model access must ensure processing speed meets business requirements.
How to set configurations
This document applies to FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
AIPROXY_API_ENDPOINT | Root address of the AI Proxy service, for example http://aiproxy:3000. Do not append /v1 or /chat/completions, and do not use a single-model service address | Required environment variable for the FastGPT App/Admin server, used to connect to the AI Proxy service |
AIPROXY_API_TOKEN | Admin Token for the AI Proxy service, must match the AI Proxy server configuration and be non-empty. Do not use application API keys, model vendor keys, or similar values | Required environment variable for the FastGPT App/Admin server, used for AI Proxy service authentication |
maxContext | 4000–8000 token | Matches the context length of tickets and chat logs |
maxTokens | 2000–3000 token | Ensures single-round reply length meets ticket processing requirements |
temperature | 0.3–0.6 | Balances answer accuracy and flexibility, suitable for problem classification and troubleshooting |
toolChoice | auto | Automatically selects tools to adapt to different ticket scenarios |
vision | Disabled | Enable on demand only if asset image upload scenarios are involved |
Three Common Misconfigurations
- Clicking model properties triggers an error, with an interface prompt of invalid configuration. Troubleshooting steps include checking the
AIPROXY_API_TOKENandAIPROXY_API_ENDPOINTenvironment variables configured on the FastGPT App/Admin server, and the authorization configuration for the corresponding model channel. - Adding
deepseek/deepseek-r1:freefails to work. The symptom is model call returns authentication failure or address error. Troubleshooting steps include checking whetherAIPROXY_API_ENDPOINTis the correct root address of the AI Proxy service, whetherAIPROXY_API_TOKENis the admin token matching the AI Proxy server configuration, and the configuration for the corresponding model channel. - Configuration is lost after Docker restart. The symptom is all model configurations and data source configurations are cleared after restart. The cause is failure to mount the FastGPT data directory to host persistent storage, resulting in container data being lost when the container is destroyed.
How to Confirm Successful Configuration
- Check that the FastGPT service startup log contains no errors about missing
AIPROXY_API_ENDPOINTorAIPROXY_API_TOKEN, or call the AI Proxy service health check interface to verify that theAIPROXY_API_ENDPOINTandAIPROXY_API_TOKENenvironment variables are configured correctly. - Submit a simulated ticket, check that the model outputs expected problem classification results, and verify that the context length configuration is reasonable.
- Restart the FastGPT service, confirm that all configurations are not lost, and verify that the persistent storage mount is working properly.
- Adjust the
temperatureparameter, check that the model's answer accuracy and flexibility meet business requirements.
The provided values are common starting points and should be measured against the reader's own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-10-19.