Deployment and upgradesOfficial documentation7 min read版本解读

FastGPT V4.8 Feature Updates and Fixes

This article records configuration and changes from a historical FastGPT release.

This article records configuration and changes from a historical FastGPT release. Use it to understand that release, and consult the release notes for your installed and target versions before applying dependencies, configuration, or migration steps.

FastGPT V4.8 Feature Updates and Fixes

Core Workflow Overhaul and New Capabilities

The FastGPT workflow engine has been fully refactored, with multiple new features and quality-of-life improvements. Key additions include:

  • Fully refactored workflow engine foundation
  • If/ElseIf/Else conditional nodes (contributed by @newfish-cmyk) for implementing branching logic; existing preview version If/Else nodes must be deleted and recreated
  • Variable update nodes (contributed by @newfish-cmyk) for modifying workflow output variables or global variables during runtime
  • Automated workflow saving and version management
  • Workflow debug mode for testing individual nodes or step-through full workflow execution
  • Scheduled app execution for cron-like recurring tasks
  • Enhanced plugin custom inputs with rendered input components
  • New share link pre-chat hook (PR #1252, contributed by @gaord)
  • 4-directional workflow connection support to simplify loop construction
  • Improved workflow context passing performance

A quick reference table for key new workflow components:

New Workflow ComponentCore PurposeRequired Action
If/ElseIf/Else Conditional NodeAdd conditional branching logicDelete and recreate existing preview If/Else nodes
Variable Update NodeModify runtime workflow or global variablesNone
Workflow Debug ModeTest individual or full workflow stepsNone

Usability, Performance, and Operational Improvements

Several updates streamline daily usage and optimize system performance:

  • Updated keyboard shortcuts: Ctrl+Enter and Alt+Enter now have improved line break cursor positioning
  • Enhanced variable configuration storage in chat sessions to prevent configuration edits from affecting active conversations
  • Simple mode now automatically updates the debug panel after configuration changes without requiring manual saves
  • Overhauled worker process management, with token calculation delegated to worker processes
  • Tool calling now supports specifying field data types: string, boolean, or number (resolved issue #1236)
  • Increased Completions API size limit (resolved issue #1241)
  • Updated Node API middleware and server-side code (contributed by @c121914yu)
  • Chat history is now trimmed to an even number of messages to support models requiring paired input, with a maximum of 50 rounds (resolved issue #1384)
  • HTTP nodes now terminate processes on error (resolved issue #1290)

Resolved Bug Fixes

All reported issues from the pre-release cycle have been addressed:

  • Fixed tool calling names being unable to start with numeric prefixes, which occurred occasionally with randomly generated IDs (contributed by @c121914yu)
  • Resolved cached global variable query parameters on share links (contributed by @c121914yu)
  • Fixed tool calling field compatibility issues (resolved issue #1253)
  • Fixed HTTP node URL cursor positioning bug (resolved issue #1334, contributed by @maquannene)

Source: FastGPT official documentation and source