Message Threading

ReedAI Assistant uses a threading model to organize conversations and maintain context across interactions. This page explains how threading works and how to effectively use it.

​

What is Threading?

In ReedAI Assistant, a thread represents a persistent conversation with an AI agent. Threads have the following characteristics:

1. Persistence: Threads are saved and can be resumed later
2. Context: Threads maintain conversation history for context
3. Organization: Threads group related messages together
4. Specialization: Threads can be associated with specific agents or tasks

​

Thread Components

A thread in ReedAI Assistant consists of:

• Thread ID: Unique identifier for the thread
• Thread Title: Descriptive name for the conversation
• Agent Type: The AI agent associated with the thread
• Created At: Timestamp of thread creation
• Updated At: Timestamp of last message
• Messages: Sequence of user messages and AI responses

​

Thread Lifecycle

Threads follow a standard lifecycle:

1. Creation: A new thread is created when starting a new conversation
2. Active Use: Messages are added to the thread during conversation
3. Archiving: Threads can be archived when no longer needed
4. Deletion: Threads can be permanently deleted

​

Context Management

Threading provides several benefits for context management:

​

Context Retention

Each thread maintains its conversation history, allowing the AI to reference previous messages. This provides:

• Continuity: Follow-up questions work naturally
• Reference: The AI can refer to previous code snippets or explanations
• Evolution: Ideas can develop over multiple exchanges

​

Context Limitations

There are some limitations to be aware of:

• Context Window: Claude 3.5 has a finite context window (approximately 200,000 tokens)
• Relevance Decay: Very long threads may have older messages that become less relevant
• Performance: Extremely long threads may impact performance

​

Thread Organization

ReedAI Assistant provides several ways to organize threads:

​

Thread Naming

Threads can be named descriptively to indicate their purpose:

• Feature-based: “User Authentication Implementation”
• Task-based: “Debugging Memory Leak”
• Project-based: “E-commerce Checkout Flow”

​

Thread Filtering

Threads can be filtered by:

• Creation date
• Last updated date
• Agent type
• Custom tags

​

Thread Archiving

Completed threads can be archived to reduce clutter while preserving the conversation.

​

Thread Best Practices

To get the most out of ReedAI Assistant’s threading system:

1. Create focused threads: Each thread should address a specific task or topic
2. Use descriptive names: Thread names should clearly indicate the purpose
3. Maintain reasonable length: Split very large tasks into multiple threads
4. Provide context: When starting a new thread, provide relevant project context
5. Archive completed threads: Keep your workspace organized

​

Thread Sharing and Collaboration

Threads can be shared with team members by:

1. Exporting: Threads can be exported as Markdown or JSON
2. Linking: Direct links to threads can be shared
3. Publishing: Threads can be published as read-only references

​

Example Thread Structure

A typical development thread might follow this pattern:

1. Initial request: User describes the feature to implement
2. Requirements clarification: AI asks questions to understand requirements
3. Implementation plan: AI proposes an approach
4. Code generation: AI provides implementation code
5. Testing strategy: AI suggests testing approach
6. Iterations: User and AI refine the implementation
7. Documentation: AI helps document the feature