This document details all resolved bugs, API adjustments, and configuration fixes for the FastGPT v4.15.0 release, for technical engineers and self-hosted operators.
Bug Fix Breakdown
| Fix ID | Issue Description |
|---|---|
| 1 | Fixed an issue where a model response error in Agent V2 mode caused steps to execute repeatedly. |
| 2 | Fixed missing charset in text responses when previewing or downloading Dataset source files. |
| 3 | Fixed abnormal default values in Workflow single-node debugging. |
| 4 | Fixed abnormal defaultConfig override behavior in model configuration. |
| 5 | Fixed TTS playback errors when adapting to the latest OpenAI SDK. |
| 6 | Fixed oversized chunks that could occur when Dataset data chunks contained code blocks. |
| 7 | Fixed abnormal multimodal file link retrieval from models. |
| 8 | Fixed potential security risks related to the indexing API, HTTP tool parsing, and private S3 object keys. |
| 9 | Fixed abnormal MCP tool expansion for tool calls after interactive nodes. |
| 10 | Fixed abnormal tool call parameter schemas for array and object types in Workflow tools. |
| 11 | Fixed UI offset in publish channel portals. |
| 12 | Fixed the v1/completions API where quoteList in nodeResponse did not return q and a. |
| 13 | Fixed session stream resume issues, including form restoration, file list restoration, node response preservation, duplicate interaction appending, temporary history titles, and cross-app chat leakage. |
| 14 | Stop-session prompts are now synchronized with the backend generation state, and the warning toast shown during stop has been removed. |
| 15 | The v1/chat/completions API previously filtered out q/a/index when returning nodeResponse; this version restores those fields. |
API Behavior Updates
Several API endpoints received targeted fixes to improve consistency:
- The v1/completions API now properly returns
qandafields within thequoteListparameter ofnodeResponseobjects. - The v1/chat/completions API restores the previously filtered
q,a, andindexfields innodeResponseresponses. - Session stream resume functionality now correctly restores form data, file lists, and node responses, prevents duplicate interaction appending, fixes temporary history title issues, and blocks cross-app chat leakage.
- Stop-session prompts now align with backend generation state, and the on-screen warning toast during stop operations has been removed.
Configuration and Tooling Fixes
All core platform and tooling issues have been fully resolved:
- Agent V2 mode no longer repeats steps following model response errors.
- Dataset source file previews and downloads now include the correct charset in text responses.
- Workflow single-node debugging uses correct default values instead of abnormal defaults.
- Model configuration
defaultConfigoverride behavior now functions as intended. - TTS playback works correctly with the latest OpenAI SDK.
- Dataset data chunking no longer produces oversized chunks for content with code blocks.
- Multimodal file link retrieval from models operates without abnormal behavior.
- Security risks related to the indexing API, HTTP tool parsing, and private S3 object keys have been mitigated.
- MCP tool expansion for tool calls after interactive nodes functions correctly.
- Workflow tool call parameter schemas for array and object types no longer exhibit abnormal behavior.
- UI offset issues in publish channel portals have been resolved.
Source: FastGPT official source
Applicability and version scope
Use this page for the documented Deployment and upgrades scenario. Confirm the FastGPT, dependency, API, and deployment versions in the official source before applying a change.
Safety guardrails
Use [REDACTED_CREDENTIAL] for credentials and private data. Confirm the documented environment and version before review.
Rollback guidance
Restore the prior technical-content authority snapshot. Restore saved configuration and data snapshots, then repeat the smallest verification scenario.