Deployment and upgradesOfficial documentation6 min readDeployment and upgrades

FastGPT 4.14.5 Stability and Performance Improvements

This document outlines the technical changes included in the FastGPT 4.14.5 self-hosted upgrade, focused on backend stability, user experience, and API re…

This document outlines the technical changes included in the FastGPT 4.14.5 self-hosted upgrade, focused on backend stability, user experience, and API reliability.

Infrastructure & Stability Optimizations

This section addresses core backend performance and reliability. Optimized Redis key retrieval logic prevents server blocking during operations that fetch large numbers of keys, reducing unexpected service interruptions. Improved reconnection logic for MongoDB, Redis, and MQ minimizes downtime by automatically restoring connections after transient failures. Additionally, the Chats database table indexes have been optimized: redundant indexes have been removed, and conditional indexes added to streamline query execution for chat-related operations.

UI & User Experience Adjustments

End-user facing refinements include support for copying content from disabled variable input fields, eliminating a previous barrier to reusing pre-configured variables. The citation list displayed below AI chat responses has been refined to only show Dataset content that was actually cited by the AI model, removing extraneous entries and improving readability of referenced source material.

LLM, API & Tooling Enhancements

This section covers improvements to large language model handling, API limits, and developer tooling. A detailed breakdown of these changes is provided in the table below:

Improvement AreaSpecific Changes
LLM Response DetectionExcludes content filter errors from being misidentified as empty AI responses, preventing false negatives when the model filters prohibited content
Error MessagingAdds more raw data to error messages for AI chat and tool calls, enabling more efficient debugging of failed requests
File Parsing APIIncreased maximum request size limit to 10MB, supporting larger file uploads for parsing
MCP SDKUpdated to the latest available version of the MCP SDK, incorporating latest bug fixes and feature improvements

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.