V4.9.1

Upgrade Guide

1. Back Up Your Database

2. Update Images

Update FastGPT image tag: v4.9.1-fix2
Update FastGPT Pro image tag: v4.9.1-fix2
Sandbox image: no update required
AIProxy image changed to: registry.cn-hangzhou.aliyuncs.com/labring/aiproxy:v0.1.3

3. Run the Upgrade Script

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/initv491' \
--header 'rootkey: {{rootkey}}' \
--header 'Content-Type: application/json'

Script Functions

Re-processes tokenization using the latest jieba dictionary. This takes a while — you can monitor progress in the logs.

New Features

Pro edition supports single-team mode for better internal member management.
Knowledge base chunk reader.
API knowledge base supports PDF enhanced parsing.
Team member invitations now use an invite link model.
Hybrid search weight configuration support.
Rerank model selection and weight configuration support. The knowledge base search weight calculation has been adjusted from vector search weight + full-text search weight + rerank weight to search weight + rerank weight. This may affect search results — you can adjust the relevant weights to adapt your data.

Improvements

Knowledge base data input UI improvements.
App-bound knowledge base data fetching moved to backend processing.
Added dependency package security version checks and upgraded some dependencies.
Model testing code improvements.
Optimized reasoning output parsing: as long as a model is configured to support reasoning, <think> tags will always be parsed, even when reasoning is disabled during a conversation.
Loaded the latest jieba dictionary for improved full-text search tokenization.

Bug Fixes

Max response tokens tooltip showing incorrect values.
HTTP Node failing to parse strings containing newline characters.
Knowledge base question optimization not passing conversation history.
Missing error message translations.
Content extraction node: incorrect schema for array types.
Model channel testing not actually targeting the specified channel.
Adding a custom model would also save default model fields, causing incorrect default model detection.
Prompt-mode tool calls not null-checking the reasoning chain, causing UI rendering errors.
Editing app info causing avatar loss.
Share link titles being reset on refresh.
Authentication failure when calculating parentPath, causing it to be cleared.