reasoning trees for AI agents

AIreasoningthatvanishesthemomentyouseetheanswer.

Branch captures every step. Navigate the tree. Fork from any node. Diff two runs. Finally understand why the answer changed.

npm install -g branch-ai
// reasoning tree · live
┌─ [root] How should I architect this?

│  ├─ Option A: monolith
│  │
│  │   └─ [fork] What if the team scales?
│  │         ├─ introduce service boundaries early
│  │         └─ refactor later (tech debt)
│  │
│  └─ Option B: services ← you are here
│        ├─ define contracts first
│        └─ deploy independently

Reasoning that vanishes
the moment you see the answer.

When an AI works through a hard problem, the reasoning disappears the moment you read the output. You got a wall of text you can't trace. If the answer is wrong, you start over. If it's right, you don't know why.

Every step. Inspectable.
Forkable. Comparable.

01
Navigate

Every reasoning step is a node. Click any point in the tree to inspect, copy, or continue from there.

02
Fork

Branch from any node. Inject a new fact mid-thought. See how the conclusion changes without starting over.

03
Compare

Diff two runs side by side. Same prompt, different assumptions — what changed?

Works where you already work.

Claude Code
OpenAI Codex
Google Gemini CLI
Factory.ai Droid

Uses your existing CLI auth. No API keys needed.

Terminal. Agent. Team.

CLI
branch 'prompt'

Captures the full reasoning tree in your terminal. Navigate it without leaving your workflow.

MCP
MCP server

Claude Code agents externalize their own reasoning. Branch intercepts and stores every step automatically.

Web
Web viewer

Share trees with your team. Public or private. Persistent links, no setup required.

Three lines to get started.

$npm install -g branch-ai
$branch "explain why this query is slow"
$branch doctor # check which CLIs are available