Open Source

10 GitHub Repos to Save 90% Tokens

Open-source tools that slash your Claude token usage — proxies, MCP servers, knowledge graphs, and smarter prompts.

01
60-90%

RTK (Rust Token Killer)

CLI proxy that filters terminal output before it hits your context

Open
  • 60-90% reduction on common dev commands
  • One binary, zero dependencies
  • Works with Claude Code, Cursor, Copilot
02
98%

Context Mode

Sandboxes raw tool output into SQLite instead of dumping it into context

Open
  • 98% context reduction on Playwright, GitHub, logs
  • Only clean summaries enter your conversation
  • Works as Claude Code plugin
03
49x

code-review-graph

Local knowledge graph that maps your codebase with Tree-sitter

Open
  • Claude reads only what matters, not the entire repo
  • 49x token reduction on large monorepos
  • 6.8x on average reviews
04
97%

Token Savior

MCP server that navigates code by symbols, not full files

Open
  • 97% reduction on code navigation
  • Persistent memory across sessions
  • 69 tools, zero external deps
05
65-75%

Caveman Claude

Makes Claude talk like a caveman to cut output tokens

Open
  • 65-75% output reduction
  • One-line install
  • Keeps full technical accuracy
06
Drop-in

claude-token-efficient

One CLAUDE.md file that keeps responses terse

Open
  • Drop-in, no code changes
  • Reduces output verbosity on heavy workflows
  • Best for output-heavy sessions
07
95%+

token-optimizer-mcp

MCP server with caching, compression, and smart tool intelligence

Open
  • 95%+ token reduction through intelligent caching
  • Compresses repeated tool outputs
08
90%

claude-token-optimizer

Reusable setup prompts for optimizing any project

Open
  • 90% token savings in 5 minutes
  • Reduces doc token usage from 11K to 1.3K
09
Ghost

token-optimizer

Finds ghost tokens that silently eat your context

Open
  • Survives compaction without losing quality
  • Fixes context quality decay
10
~40%

claude-context (by Zilliz)

Code search MCP that makes your entire codebase the context

Open
  • ~40% reduction with equivalent retrieval quality
  • Hybrid BM25 + dense vector search

How to Stack Them

You don't need all 10. Pick 2-3 based on your workflow:

Heavy terminal output?

RTK

Big codebase?

code-review-graph + Token Savior

Lots of MCP servers?

Context Mode

Quick fix?

Caveman + claude-token-efficient

Run /context in a fresh session

Most people are burning tokens without knowing it. See how much is gone before you even type a word — your pocket will thank you later.