CherryHQ/cherry-studio

Knowledge base retrieval logic regression in v1.7.1 affecting novel writing continuity

Open

#11626 opened on Dec 1, 2025

View on GitHub
 (2 comments) (1 reaction) (1 assignee)TypeScript (45,688 stars) (4,338 forks)batch import
discussionhelp wanted

Description

[!NOTE] This issue/comment/review was translated by Claude.

Issue Checklist

  • I understand that issues are meant for feedback and problem-solving, not for venting, and I will provide as much detail as possible to help resolve the issue.
  • I have checked the pinned issues and searched through the existing open issues, closed issues, and discussions and did not find a similar suggestion.
  • I confirm that I am here to ask questions and discuss issues, not to report bugs or request features.

Platform

Windows

Version

1.7.1

Your Question

I am using AI to write novels and encountered a problem. My article reached the maximum context of a single conversation when writing chapter 8. For this reason, I started a new conversation and imported the old conversation into the knowledge base to help the AI better connect the content when writing. However, I encountered a very serious problem. The logic used by Cherry Studio version 1.7.1 is: first let the AI think → AI decides whether to retrieve the knowledge base → AI outputs content. But when continuing to write novels, this logic is actually harmful. In order to make the AI output content that can connect with the previous conversation, I spent an extra 10 times to correct this AI. If it were the more advanced knowledge base retrieval logic of version 1.5, the AI would retrieve the knowledge base first after I issued the instruction, then think with the content retrieved from the knowledge base, and then output. But in version 1.7.1, the AI will not do this at all. I don't understand how to make the AI successfully complete cross-conversation plot connections, and I don't understand why developers would have such a regression.

Context

No response

Additional Information

Priority

Low (Review when available)

Issue Checklist

  • I understand that issues are meant for feedback and problem-solving, not for venting, and I will provide as much detail as possible to help resolve the issue.
  • I have checked the pinned issues and searched through the existing open issues, closed issues, and discussions and did not find a similar suggestion.
  • I confirm that I am here to ask questions and discuss issues, not to report bugs or request features.

Platform

Windows

Version

1.7.1

Your Question

我正在用AI写小说，我遇到了一个问题，我的文章在写到第八章时达到了单个对话最大上下文，为此我新开了一个对话并将旧对话导入知识库以便于AI写作时能够更好衔接，但我遇到了一个非常严重的问题，1.7.1版本的Cherry studio使用逻辑是先让ai思考→ai自主决定是否检索知识库→ai输出内容，但在续写小说时这套逻辑反而是有害的，为了让ai输出一个能和上一个对话里衔接上的内容，我又额外花了10次去纠正这个ai。如果是1.5版本更先进的知识库检索逻辑，ai会在我发出指令后先检索知识库，然后带着知识库检索到的内容思考，然后再输出，但在1.7.1版本ai完全不会这么做。我不明白如何让ai顺利完成跨对话剧情衔接，我也不理解为什么开发者会出现这种退步

Context

No response

Additional Information

Priority

Low (Review when available)

Contributor guide

Tech stack
typescript
Domain
aidesktopbackend
Issue type
bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
needs maintainer response
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
needs investigation
Prerequisites
TypeScriptknowledge base retrieval logic
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
20
Research direction
Investigate the regression in knowledge base retrieval logic between Cherry Studio versions 1.5 and 1.7.1. Focus on the order of operations: in v1.5, the AI retrieves the knowledge base before thinking, while in v1.7.1, it thinks first and then decides whether to retrieve. Look at the relevant source code files responsible for conversation context and knowledge base integration, such as those related to AI prompts or retrieval pipeline. Check recent commits or changelogs for changes that may have altered this behavior. The issue is filed as a question, so maintainer clarification is needed on whether this is a bug or an intentional design change.