Autonomous coding agents are powerful, but they often come with a steep token tax. Tools like Claude Code can rapidly consume your usage limits during reviews by rescanning projects, pulling broad context, and running background operations.
AI Badger offers a better way: map your codebase locally, extract only the precise snippets needed, and conduct thoughtful reviews inside your favorite web chat.
Two Ways You Save Tokens
- Dramatically more efficient usage — Badger eliminates noisy full-repo scans and recursive context loops. The AI receives clean, surgical context instead of everything.
- Leverage fixed-cost web chats — Most web interfaces run on flat-rate monthly subscriptions (like ChatGPT Plus or Claude Pro) or offer generous free daily tiers. By offloading heavy discussions to the web chat, you leverage a fixed-cost window and avoid burning through your metered API balances or rolling agent limits.
Even on platforms where limits are shared, Badger helps you stay efficient and avoid the faster-burning autonomous modes.
Best Code Reviews with badger review
AI Badger shines in review mode. Here's a complete walkthrough using a simple React todo app.
Step 1: Launch Review Mode
cd my-react-app
badger review
Badger pre-loads your git diff with a default prompt. Edit it before submitting:
Review the following change for concrete bugs, edge cases, maintainability issues, and
unintended behavior changes. Focus on issues I should fix before committing.
Step 2: Copy Prompt 1 (Compact Map)
Badger outputs a lightweight project topology (structure, key files, your diff summary). Paste into Claude Web, ChatGPT, DeepSeek, or any other chat.
Step 3: Targeted Extraction
The AI replies with precise selectors:
FILE:src/App.tsx
NEAR:src/App.tsx#function handleSubmit
PREFIX:src/components/TodoItem.tsx#return
Paste them back into Badger. It extracts only the relevant code.
Step 4: Paste Focused Prompt 2
Badger assembles clean context. Paste into your web chat — get high-quality review feedback without token waste.
Step 5: Apply Changes
Paste the AI's suggestions back into Badger. Review the write plan and apply fixes with confidence.
Trivial React Example
Consider a small change in a todo app:
// Missing key + potential re-render issues
{todos.map(todo => <TodoItem todo={todo} />)}
Why this matters (even for non-React devs): When rendering lists in React, each item should have a unique key. Without it, the UI can behave strangely when items are added or removed. A good code review also flags small performance and accessibility improvements.
Badger + web chat quickly flags:
- Missing
keyprop - Opportunities for
React.memooruseCallback - Accessibility improvements
All while using far fewer tokens than a full autonomous agent.
Why This Workflow Wins
- Precision-first — Only requested files/spans are sent
- Fixed-cost advantage — Maximize value from subscriptions like ChatGPT Plus and Claude Pro
- Review-optimized — Focuses the AI on risks and correctness
- Privacy & control — Fully local until you copy; explicit consent for writes
- Universal — Works with React, Go, Java, Node.js, Python, etc.
Install and Try It
brew tap pvrlabs/aibadger
brew install pvrlabs/aibadger/badger
Or:
curl -fsSL https://raw.githubusercontent.com/PVRLabs/aibadger/main/install.sh | sh
Run badger review in any repo with changes.
Full documentation: docs/usage.md