tirth8205/code-review-graph

やべ、リンク飛べなくなってた メモ覗き見した人こっちからどうぞ ①Caveman github.com/juliusbrussee/cav… ②Token Savior github.com/mibayy/token-savi… ③RTK github.com/rtk-ai/rtk ④Code Review Graph github.com/tirth8205/code-re… ⑤Claude Token Optimizer github.com/nadimtuhin/claude… ⑥Token Optimizer github.com/alexgreensh/token… ⑦Token Optimizer MCP github.com/ooples/token-opti… ⑧Claude Context github.com/zilliztech/claude… ⑨Context Mode github.com/mksglu/context-mo… ⑩Claude Token Efficient github.com/drona23/claude-to…

Matched through a repo-specific project phrase query.

Use this one command to reduce your claude code token usage by 27X times. Al coding tools like claude re-read your entire codebase on every task which shoot up your token usage. code-review-graph fixes that. It builds a structural map of your code with Tree-sitter, tracks changes incrementally, and gives your Al assistant precise context via MCP so it reads only what matters.

Matched through a repo-specific project phrase query.

一个叫 code-review-graph 的工具通过调用链分析，把上下文从27,732 个文件压到 15 个。 49 倍压缩。不是砍掉有用信息，是只保留真正相关的部分。 原理不复杂：Tree-sitter 解析 → 构建图谱 → 追踪爆炸半径 → 只喂必要上下文。 code-review-graph 做的事可以用一句话说清楚： 把"AI 读整个仓库"变成"AI 只读爆炸半径内的文件"。 你改了 login()，它追踪所有调用 login 的函数、所有依赖 login 的模块、所有测试 login 的用例。 AI 拿到的上下文是：这个改动影响了什么，哪些测试可能挂，风险等级是多少。 不需要读 27,000 个文件来回答这个问题。 本地 SQLite 存储，不传云端。

Matched through a repo-specific project phrase query.

Everyone’s chasing bigger context windows, but the smarter move is burning 80% fewer tokens like this list suggests. Most of these tools (especially Caveman, RTK, and Code Review Graph) prove that smarter context > more context. The real flex isn’t paying for Opus 4.7 with massive windows — it’s staying efficient so you can run longer sessions without the bill shocking you. 2026 winners will optimize tokens ruthlessly, not just scale them.

Matched through a repo-specific project phrase query.

10 github repos to cut your claude code tokens by 60-90%. rtk, context mode, code-review-graph, token savior, caveman, token-optimizer-mcp, claude-context, and more. all verified, all free. plus which 2-3 to actually stack. aiia.ro/blog/claude-code-tok…

Matched through a repo-specific project phrase query.

3. code-review-graph 使用 Tree-sitter 在本地构建代码库知识图谱 Claude 只读真正相关的内容，而不是整个仓库 在大型单体仓库中可减少 49 倍 token 消耗 平均代码审查可减少 6.8 倍 仓库： github.com/tirth8205/code-re…

Matched through a repo-specific project phrase query.

If you're using Claude Code / Codex, it’s likely re-reading your whole repo every time. That’s expensive + inefficient. code-review-graph changes this: - builds AST → dependency graph - tracks changes incrementally - computes “blast radius” - sends only relevant files to LLM Context selection > prompt engineering. Repo in comments.

Matched through a repo-specific project phrase query.

🐍 code-review-graph ⭐ 199 stars Graphe de connaissances pour la revue de code qui économise vos tokens IA #GitHub

Matched through a repo-specific project phrase query.

【衝撃のトークン削減】code-review-graphがすごすぎる。 Claude Code等でコードレビューする際、プロジェクト全体を読まずに「影響範囲」だけを特定してAIに渡すローカル知識グラフ構築ツール。 レビュー時のトークン消費を最大80%削減。大規模リポジトリのAI開発が劇的に効率化されます。

Matched through a repo-specific project phrase query.