Reduce Claude Desktop/Web token usage by up to 27x with intelligent structural code mapping and on-demand context fetching. No API hacks. No configuration. Just pure efficiency.
Interactive graph showing relationships between files, classes, functions, and tests. Query precisely instead of reading everything.
Traditional approach vs. Context Optimizer. The results speak for themselves.
Prompt-native tooling that works within Claude's official features.
Eliminate full-file dumps with fetch-on-demand workflows and structural reasoning over raw text scanning.
See your codebase as a connected graph. Understand relationships, dependencies, and impact instantly.
Know exactly what breaks when you change something. Trace callers, dependents, and affected tests.
Persistent workflow that applies to all Claude sessions. Set once, benefit forever across all chats.
Instant token-saving mode for single chats. Copy-paste and start optimizing immediately.
Lightweight manifest generator extracts functions, classes, and relationships — no heavy dependencies.
No leaked code. No reverse engineering. No API hacks. Built from scratch using only official Claude features.
Python, TypeScript, JavaScript, Go, Rust, Java, C++, Ruby, PHP, Swift, Kotlin, and more.
Adjust token budgets, file limits, snippet lengths, and risk scoring to match your workflow.
Three simple steps to massive token savings.
Get up and running in three steps:
# Step 1: Clone and generate manifest
git clone https://github.com/anshmajumdar121/context-optimizer
cd context-optimizer
python tools/context_mapper.py /path/to/your/project
# Step 2: Add to Claude Custom Instructions
Copy skill/claude-custom-instructions.md → Claude Settings
# Step 3: Start coding efficiently
Paste prompt/one-click-vertical-prompt.md in your chat
✅ Ready! You're now using 5-27x fewer tokens.
Join developers saving 5–27x on Claude token usage. Open source, free forever, works today.
MIT Licensed · No API keys required · Works offline