V4.8.9 (Requires Initialization)

Upgrade Guide

1. Back up your database

2. Update images

• Update the FastGPT image tag to v4.8.9
• Update the FastGPT commercial edition image tag to v4.8.9
• Sandbox image update is optional

3. Run initialization (Commercial Edition)

From any terminal, send an HTTP request. Replace {{rootkey}} with the rootkey from your environment variables, and {{host}} with your FastGPT domain:

curl --location --request POST 'https://{{host}}/api/admin/init/489' \
--header 'rootkey: {{rootkey}}' \
--header 'Content-Type: application/json'

This initializes notification methods for multi-tenancy (internal use only, no action needed).

V4.8.9 Release Notes

1. New - File upload configuration. Image upload capability is now determined by system configuration rather than relying on vision model availability.
2. New - AI chat node and tool calls support toggling "Enable image recognition". When enabled, it automatically retrieves images uploaded in the chat and image URLs from "User question".
3. New - Document parsing node.
4. Commercial - Team notification account binding for receiving important messages.
5. Commercial - Knowledge Base collection tagging for tag-based management.
6. Commercial - Knowledge Base search node supports tag filtering and creation date filtering.
7. Commercial - Transfer app owner permissions.
8. New - Delete all conversation starters.
9. New - QA splitting supports custom chunk sizes, and optimized the issue where GPT-4o-mini produced very little content with large chunks.
10. Improved - Lazy loading for chat messages to reduce network transfer.
11. Improved - Clear file selection cache to allow re-selecting the same file.
12. Fixed - Knowledge Base file upload progress not reaching 100% under unstable network or with many files.
13. Fixed - After deleting an app, returning to chat and selecting the last conversation from the deleted app showed an error.
14. Fixed - Plugin dynamic variable default values not displaying correctly.
15. Fixed - Tool call temperature and max response values not taking effect.
16. Fixed - In function call mode, GPT models require the content parameter in assistant role messages (doesn't affect most models since nearly all have switched to ToolChoice mode; FC mode is deprecated).
17. Fixed - Knowledge Base file upload progress updates could be incorrect.
18. Fixed - Knowledge Base page always resetting to the first page during rebuilding.
19. Fixed - Knowledge Base list OpenAPI authentication issue.
20. Fixed - Unable to provide feedback on new conversations via share links.