Feedback loops for
[ engineers ] &&
[ their agents ]

What is Branchmore?

Branchmore is an analytics platform for engineers using AI coding tools like Claude Code, Codex and Cursor. It gives you a clear picture of how you use AI: what it costs, how much you're actually getting done with it, and whether it's behaving safely.

Rather than flying blind on AI spend and activity, Branchmore centralizes your usage data so you can make informed decisions about your costs, your productivity, and your security.

What can I
 get out
 of
 Branchmore?

See what you're spending

Track your AI subscription costs alongside token usage so you know whether your plan is worth it and when you might be able to downgrade.

Measure your output

Correlate your AI usage with GitHub commit activity to understand whether your AI tools are translating into real code shipped.

Stay safe

Review your session transcripts, spot where you and your AI disagreed, and catch risky tool calls before small mistakes become big problems.

Stop hitting rate limits

See when and why you're getting throttled, and use that data to right-size your plan.

How do I get started?

Getting useful data into
Branchmore takes about
10 minutes:

Install the bmor CLI

This runs alongside Claude Code and Codex to capture your session data and
forward it to Branchmore. See the
Getting Started page for installation options
(curl, Homebrew, or npm).

Connect GitHub

Links your commit activity so you can see how your AI usage maps
to code you actually shipped.

Add your SaaS subscriptions

Enter your recurring AI costs (Claude, Cursor, etc.) so Branchmore can show
you real
cost-per-output numbers.

The Getting Started page tracks
your progress
through these
steps automatically.

How do I leave feedback?

We genuinely want to hear from you. Feedback helps us prioritize what matters most. Use the Feedback page to share thoughts, feature requests, or anything on your mind. If you need help with something specific, visit the Support page.

The Branchmore team
Andrew Shu, Branchmore founder

Hey! My name is Andrew Shu, Branchmore's founder. I'm a software engineer and engineering leader with 16+ years of experience at Permiso Security, Cisco Meraki, and Printivity, building impactful, secure products for scrappy startups and large enterprises alike.

As I bootstrap Branchmore, I'm building with Anes Sprecic, a talented senior engineer who contracts through Elixirator. Anes brings 7+ years of experience building communication software at Bicom, and enterprise dealership platforms at Elixirator.

Why build Branchmore?

While leading a team at Permiso, I helped them build and adopt AI coding tools, and saw how much subtle work it took to actually scale their productivity. Tuning context, picking the right tool for the job, keeping spend in check, knowing when the AI was helping versus quietly making things worse. Every engineer wanted to get better at AI, but few had the time to dig into it.

Branchmore gives you clarity into how AI is being used on your codebase: what it costs, what it produces, and the safety risks it's taking on your behalf. And hopefully, have a little fun with AI!

We're early in our journey. Right now we're focused on collecting, organizing, and visualizing your data. Over the next few months we'll turn that into actionable insight: where your bottlenecks are, and what's dragging it down.

I'd love to chat! Your ideas and feedback shape what we build; tell me what you're trying to do with your AI tooling. You can reach me directly on X, LinkedIn or my blog: