Skip to content

[Feature Request] List of Message compression by using LLM doing summarize #624

@Wendong-Fan

Description

@Wendong-Fan

Required prerequisites

Motivation

When dealing with large volumes of message information, we need a way to condense conversations while preserving their meaning. Using LLMs for summarization allows us to maintain a structured format that keeps key information intact, following the schema outlined in issue #591.

Solution
Create a MessageSummarizer utility that:

Takes CAMEL messages as input
Produces structured summaries that preserve essential information
Maintains compatibility with existing ChatAgent and Message implementations
The summarizer will work similarly to how ChatAgent handles messages, but with a focus on condensing content while retaining critical context and labels.

Solution

No response

Alternatives

No response

Additional context

No response

Metadata

Metadata

Assignees

Labels

Type

No type

Projects

Status

Sprint Planned

Milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions