What this category's data looks like
Data for this category comes from material grades, customer technical requirements, temperature control instruments, furnace temperature uniformity tests, hardness testers and metallographic equipment, and MES systems. The update frequency is per heat and per batch. Daily documents include heat treatment process cards, charging charts, furnace temperature curves, hardness and metallurgical reports, equipment inspection forms, and exception handling sheets. Each document has dedicated fields: process cards mark heating temperature (℃) and holding time (minutes). Furnace temperature curves record temperature data with a 10-second sampling interval. Hardness reports mark mechanical performance values in HRC or HV units.
What constraints these characteristics impose on the tool calling and plugins link
The high-frequency update nature of this category’s data requires tool calling to pull latest furnace temperature curves, hardness reports and other data in real time, to avoid using expired information. The mixed scenario of multiple document types requires plugins to support parsing different formatted content such as tables, curve text, and revision annotations. Data sources involve workshop local equipment and cloud systems, so tool calling must adapt to cross-environment permission configurations. The serious consequences of errors require tool calling to add parameter verification and failure retry mechanisms to ensure the accuracy of process judgment. The existence of dedicated fields requires tool calling to match specific fields such as material grades and furnace parameters to avoid invalid queries. At the same time, tool calling return results must be limited to 10 entries or fewer to avoid overload.
How to set the configuration
This page is compatible with FastGPT v4.17.0. The following are specific configuration instructions:
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
| Workflow/API tool selection | auto | Adapts to parsing needs for multiple document types including heat treatment process cards and furnace temperature curves |
| Global variable passing configuration | Bind workflow global token field | Meets cross-tool parameter transfer requirements |
| Local MCP call permission | Enable local service access whitelist | Adapts to workshop local equipment data pulling |
| Knowledge base associated query settings | Calibrate based on actual measurements | Matches the dedicated field matching accuracy of heat treatment documents |
| Word document parsing plugin | Enable revision content extraction module | Handles revision typesetting requirements for process cards |
| Workflow/API function-call configuration | Enable mandatory parameter verification | Prevents abnormal data from entering the process judgment flow |
Three common misconfigurations
- When calling MCP tools in a workflow, the global variable token is empty, and the interface displays a parameter missing error. The cause is that the binding relationship between the global variable and the tool input parameters is not configured.
- Calling the local MCP tool fails, returning connection timeout or permission denied. The cause is that the local service access whitelist is not enabled.
- Associated knowledge base query returns no relevant content, with the result count being 0. The cause is that dedicated fields of heat treatment documents (such as material grades and furnace parameters) are not matched.
How to confirm the configuration is complete
- Trigger a test workflow, check if the MCP tool input parameters carry the configured global token field, and confirm the parameter is not empty.
- Try calling the local MCP tool, check if the return result is normal with no connection timeout or permission errors.
- Initiate a query associated with the heat treatment knowledge base, check if the return result includes dedicated document content such as process cards and furnace temperature curves.
- Upload a Word process card with revisions, check if the parsing result includes revision mark content.
The values given are common starting points and should be measured against the reader's own samples.