claude-token-analyzer
About
Diagnoses token waste in Claude Code sessions with 6 anomaly types and severity scoring. Fully local.
Details
- Author
- li195111
- Categories
- Developer Tools, AI, Other
Jump to
Setup
Install claude-token-analyzer in your MCP client (Claude Desktop, Cursor, Windsurf, and others).
Repository: https://github.com/li195111/claude-token-analyzer
Follow the installation instructions in the repository README, then restart your MCP client.
Diagnoses token waste in Claude Code sessions with 6 anomaly types and severity scoring. Fully local.
Your Claude Code sessions might be burning tokens you can't see.Diagnoseswhere your tokens go, why they're wasted, and what to fix first.
Fully local— parses your~/.claudeJSONL files into SQLite. Nothing leaves your machine. No cloud. No telemetry.
- Diagnose token waste— Detects 6 statistical anomaly types (HighCost, LowCacheHitRate, CostInefficient, ExcessiveToolUse, HighTokenUsage, UnusualModelMix) with severity scoring
- Audit costs— Per-session, per-project, and global cost breakdowns with monthly trends
- Forecast spending— Daily/weekly/monthly usage trends with burn-rate projections
- Optimize cache— Identifies sessions with poor cache hit rates that inflate costs
- Prioritize fixes— Severity-scored anomalies so you know what to fix first, not just what's wrong
- Converse naturally— Ask in plain language: "how much did I spend?" or "scan for anomalies"
# Install (binary auto-downloads, no Rust toolchain needed) claude plugin install claude-token-analyzer
Then just ask in any Claude Code session:
> cta > how much did I spend this month? > scan for anomalies > analyze this project > show me usage trends
~/.claude/projects//.jsonl Your session logs (never modified) → parser.rs Extract + deduplicate responses → analyzer.rs Cost calculation, 10-dimension metrics → storage.rs Upsert into local SQLite → detector.rs 6-type anomaly detection + severity scoring → MCP tools / Skills You ask, it answers
All processing happens locally. The SQLite database lives under the plugin data directory in plugin mode and falls back to~/.claude/in standalone mode. No network calls, no external dependencies at runtime.
- DB/archive: environment variable > plugin mode ($CLAUDE_PLUGIN_ROOT) > standalone mode ($HOME/.claude/)
- Projects: environment variable > Claude config dir ($CLAUDE_CONFIG_DIR/projects) > standalone mode ($HOME/.claude/projects)
git clone https://github.com/li195111/claude-token-analyzer.git cd claude-token-analyzer bash scripts/build.sh # Binary: mcp-server/target/release/cta-mcp-server # Run tests (106 tests) cargo test --all-targets --manifest-path mcp-server/Cargo.toml # Lint cargo clippy --manifest-path mcp-server/Cargo.toml -- -D warnings # Launch with plugin loaded claude --plugin-dir .
Issues and PRs welcome! Seeopen issuesfor starter tasks.
- Clone the repo and runbash scripts/build.sh
- Runcargo test --all-targets --manifest-path mcp-server/Cargo.tomlto verify
- Load the plugin locally withclaude --plugin-dir .
Rust toolchain required. The project usescargo clippy -- -D warningsfor linting.
你的 Claude Code 會話可能正在浪費你看不見的 token。診斷token 流向、浪費原因，並告訴你該優先修正什麼。
全本地運行— 解析~/.claudeJSONL 檔案到 SQLite。資料不離開你的機器。無雲端、無遙測。
- 診斷 token 浪費— 6 種統計異常類型，含嚴重度評分
- 成本審計— 按會話、專案、全域的費用拆解與月度趨勢
- 趨勢預測— 日/週/月用量趨勢與燃燒率預測
- 快取優化— 識別低快取命中率的會話，降低不必要開銷
- 嚴重度排序— 優先處理影響最大的問題，不只是標記異常
- 自然語言互動*— 用中文直接問：「看看狀況」「這個月花多少」「有異常嗎」
# 安裝（自動下載 binary，無需 Rust 工具鏈） claude plugin install claude-token-analyzer
> 看看狀況 > 這個月花多少？ > 有異常嗎？ > 分析這個專案 > 用量趨勢
This is a web browser that enables your coding agent, such as Claude Code, to visit websites on your behalf and assist you in identifying bugs or creating UI test cases.
A stateful LSP runtime for AI agents: warm language server sessions with 50+ tools for go-to-definition, find-references, diagnostics, rename, and more across 30+ languages.
AI-powered code quality analysis to detect best practice violations, security issues, and architectural problems in real-time.
Automatically generates documentation for code repositories by analyzing directory structures and code files using the OpenRouter API.
MCP server for the Brainfuck programming language that allows your favourite LLM to debug Brainfuck programs.
Provides fast C++ code intelligence for LLMs using the clangd language server.
A server for code modification and generation using Large Language Models.
Your AI Code Review Council - Get diverse perspectives from multiple AI models in parallel.
MCP server that gives coding agents program-analysis primitives — data flow, call graphs, taint analysis — so they reason from ground truth instead of grep-and-guess. (same as the GitHub About — keeps your messaging consistent across the web).
CLI token optimizer and AI context generator with built-in MCP server. Scans codebases to extract routes, schema, components, and dependencies 9x–13x token reduction for Claude Code, Cursor, Copilot, Codex, and Windsurf.
An MCP server for the codetoprompt library, enabling integration with LLM agents.
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
