User Experience Refinements
This release includes multiple UI and workflow improvements to reduce user friction:
- Mutually exclusive parent and child node selection: Eliminates unexpected UI jitter that occurred when selecting and moving both parent and child nodes simultaneously within the application’s knowledge tree.
- File injection context relocation: Moved message file injection logic from the system message context to the user message context, which increases cache hit rates for repeated, similar prompt sequences.
- Balance error messaging updates: Revised insufficient-balance notification text for non-administrator users and guest visitors to deliver clearer, more consistent feedback on quota limitations.
- Template visibility controls: Hides template selection options when the authenticated user lacks the required permission to create new resources, reducing visual clutter and confusion.
A new configurable option was added to improve model compatibility, as detailed below:
| Configurable Toggle | Purpose |
|---|---|
| Convert images to base64 | Encode uploaded images as base64 strings prior to transmission to AI models |
Security Hardening Improvements
Critical security updates address potential exploit vectors:
- Third-party dataset SSRF protection: Enhanced security controls for all requests made to third-party dataset sources to mitigate server-side request forgery (SSRF) risks, preventing unauthorized access to internal network resources.
- Codex sandbox AST validation: Strengthened abstract syntax tree (AST) inspection within the codex-sandbox execution environment to block known code execution bypass attempts for untrusted user inputs.
- IP validation strengthening: Updated IP address validation logic across public and internal endpoints to prevent spoofing-based access bypasses, tightening access control for both authenticated and unauthenticated users.
Operational Stability Fixes
Several stability and alerting improvements reduce operational overhead for self-hosted deployments:
- Duplicate rate limit alert suppression: Prevented duplicate site synchronization rate-limit messages from appearing in admin logs and user notifications, cutting down on unnecessary alert noise during repeated sync operations.
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.
Rollback guidance
Restore the prior technical-content authority snapshot. Restore saved configuration and data snapshots, then repeat the smallest verification scenario.