Feature-Request: The One Feature That Would Completely Change How People Use ChatGPT
OpenAI Developer Community
June 20, 2026
Feature Request: Conversation Branching, Side Threads, and Tree-Based Discussions
Summary
I would like ChatGPT conversations to support branching discussion trees, allowing users to start independent side conversations from any specific sentence, paragraph, message, or response without disrupting the main discussion.
Rather than treating conversations as a single linear chain, ChatGPT could optionally function as a structured conversation tree where topics naturally branch into related subtopics.
The Problem
Human curiosity is rarely linear.
While discussing one topic, ChatGPT often mentions several related ideas that deserve exploration. However, asking about those ideas typically redirects the main conversation away from its original purpose.
For example:
Main Topic: Frederick the Great
ChatGPT mentions:
- Military reforms
- Battlefield tactics
- The Seven Years’ War
- Enlightenment philosophy
If I ask about battlefield tactics, the conversation may drift away from Frederick himself. If I then ask about Enlightenment philosophy, it drifts further still.
To avoid this, many users create multiple chats, open multiple tabs, use multiple devices, or continually start new conversations. This creates fragmented context and makes knowledge exploration less efficient.
Proposed Solution
Allow users to create a branch from:
- Any sentence
- Any paragraph
- Any message
- Any user message
- Any ChatGPT response
The user could highlight text and select:
- Branch Conversation
- Open Side Thread
- Explore This Topic
- Ask Follow-Up Here
This would create a linked sub-conversation while preserving the original discussion.
Example Workflow
Main Conversation:
«Frederick the Great developed innovative battlefield tactics during the Seven Years’ War.»
I highlight:
«“innovative battlefield tactics”»
and ask:
«Which tactics specifically?»
A side thread opens.
Within that thread, ChatGPT explains oblique order tactics.
I then highlight:
«“oblique order”»
and ask:
«Who invented it?»
A second-level branch opens.
The original Frederick discussion remains unchanged and easy to return to.
Conversation Tree Structure
Instead of:
Conversation A
becoming:
Conversation A → Topic B → Topic C → Topic D
the structure becomes:
Conversation A
├── Topic B
│ ├── Topic B1
│ └── Topic B2
├── Topic C
│ ├── Topic C1
│ └── Topic C2
└── Topic D
This more accurately reflects how people learn and investigate information.
Key Benefits
Reduced Topic Drift
The main discussion remains focused.
Better Research
Users can investigate details without losing context.
Improved Learning
Complex subjects naturally break into manageable subtopics.
Less Tab Clutter
Users no longer need multiple chats, browser tabs, or devices to manage related questions.
Better Context Preservation
The AI retains the local context relevant to each branch rather than forcing unrelated topics into one conversation stream.
Enhanced Knowledge Discovery
Users can follow curiosity wherever it leads while maintaining organization.
Optional Features
Branch Summaries
Automatically generate a summary for each branch.
Branch Naming
Allow users to rename branches manually.
Branch Search
Search across the entire conversation tree.
Branch Merging
Merge useful discoveries back into the parent conversation.
Branch Comparison
Compare two branches side-by-side.
Branch Bookmarks
Mark important branches for later review.
Collapsible Tree View
Expand and collapse branches to reduce visual clutter.
Branch Export
Export individual branches as separate chats, documents, or research notes.
Advanced Concept: Knowledge Map Mode
An optional interface could display conversations as an interactive knowledge graph.
Topics would appear as connected nodes rather than a simple message timeline.
Users could visually navigate relationships between ideas and jump directly to any branch.
This would transform ChatGPT from a linear chatbot into a structured exploration and research environment.
Why This Matters
Many users already simulate this feature by opening multiple chats, browser tabs, windows, or devices to preserve discussion structure. A native branching system would formalize an existing user behavior, reduce friction, improve organization, and enable much deeper exploration of complex topics.
Conversation branching would make ChatGPT feel less like a chat log and more like a living knowledge tree that grows naturally alongside the user’s curiosity.
Discussion in the ATmosphere