#406 · Primary category: AI Coding Assistants
token-optimizer-mcp
Measure token savings per AI coding agent, optimize context, and share a live local knowledge graph across 16 CLI clients.
Project last updated:08/30/26
GitHub Stars
501
Forks
59
Contributors
7
License
MIT
Why we included this project
Coding agents often spend their largest context expenses on work they already did: re-reading files that haven't changed, dumping a whole file to see a few lines, rerunning broad searches, and rediscovering conclusions from an earlier session. Token Optimizer is an MCP server that sits between a coding client and those built-in calls, refuses the expensive ones, and hands back a cached, diff-based replacement instead. It works with Codex, Claude Code, Gemini, and other CLI clients (sixteen in total), and it keeps a per-project knowledge graph of findings and decisions, so when an agent touches a related file the answer comes back almost immediately at a fraction of the original token cost. The dashboard also attributes savings to each client and flags the cases where the optimizer itself costs more than it saves. For teams running agents on large codebases, that is a practical way to reclaim context headroom without touching your prompts.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
superpowers
An agentic skills framework & software development methodology that works.
ECC
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
skills
Skills for Real Engineers. Straight from my .agents directory.
opencode
The open source coding agent.
andrej-karpathy-skills
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.