Token efficiency: tools for developers

Tools that reduce token costs, compress context or cache LLM requests. An entry can appear under several use cases when its topics or description provide several explicit signals.

Use case

Activity

Sort by

101 entries in this view.

Tool ranking

Ranked by known GitHub stars, highest first; archived repositories come after maintained repositories. Raw values keep their source-specific window, shown under each tool; a change requires at least two snapshots from the last 7 days.

  1. 1

    CLI
    Active

    CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies

    RustLicense: Apache-2.0
    • agentic-coding
    • ai-coding
    • anthropic
    • claude-code
    • cli

    githubmeasured growthOpen source ↗

    78 098GitHub stars+671 (+0.87 %)
  2. 2

    MCP
    Active

    Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.

    PythonLicense: Apache-2.0
    • agent
    • ai
    • anthropic
    • claude-code
    • compression

    githubmeasured growthOpen source ↗

    68 221GitHub stars+619 (+0.92 %)
  3. 3

    Other
    Active

    Find the ghost tokens. Fix them. Survive compaction. Avoid context quality decay.

    Python
    • agentskills
    • claude-code
    • claude-code-skill
    • claude-plugin
    • context-engineering

    githubmeasured growthOpen source ↗

    2 084GitHub stars+120 (+6.1 %)
  4. 4

    Other
    Active

    Open-source TypeScript terminal coding agent for DeepSeek-V4 — builds on DeepSeek's strong price-performance and ultra-cheap cache pricing, engineering byte-stable prefixes and cache-reusing forks so cross-session memory and a continuous…

    TypeScriptLicense: MIT
    • ai-agent
    • cli
    • coding-agent
    • deepseek
    • llm

    githubmeasured growthOpen source ↗

    1 254GitHub starsstable
  5. 5

    MCP
    Active

    Pre-injection context compression for coding agents. Zero LLM calls, zero telemetry, offline-safe.

    RustLicense: NOASSERTION

      mcpmeasured growthOpen source ↗

      609GitHub starsstable
    • 6

      Skill
      Dormant

      Give Claude Code photographic memory in ONE portable file. No database, no SQLite, no ChromaDB - just a single .mv2 file you can git commit, scp, or share. Native Rust core with sub-ms operations.

      TypeScriptLicense: MIT
      • ai-tools
      • anthropic
      • claude
      • claude-agents
      • claude-ai

      githubmeasured growthOpen source ↗

      Install git clone https://github.com/memvid/claude-brain ~/.claude/skills/claude-brain

      574GitHub starsstable
    • 7

      MCP
      Active

      Context engineering for AI agents. ~80% fewer tokens. Fix tool overload. Skills and memory with in-process BM25 and semantic retrieval. Progressive Disclosure. No vector DB.

      TypeScriptLicense: MIT
      • accuracy
      • agents
      • claude-skills
      • context
      • harness

      githubmeasured growthOpen source ↗

      430GitHub stars+3 (+0.70 %)
    • 8

      Skill
      Active

      23 Claude Code plugins: TDD enforcement hooks, git/PR workflows, spec-driven development, code review, project lifecycle, fix-from-error, maintenance automation, context optimization, research, and multi-LLM delegation. 186 skills, 128…

      PythonLicense: MIT
      • ai-agents
      • architecture-patterns
      • awesome-claude-code
      • awesome-claude-skills
      • claude-agents

      githubmeasured growthOpen source ↗

      Install git clone https://github.com/athola/claude-night-market ~/.claude/skills/claude-night-market

      332GitHub stars+2 (+0.61 %)
    • 9

      MCP
      Active

      MCP server and proxy that compresses LLM prompts, tool output, and replies to cut token cost.

      ↓ 171 downloads/weekRustLicense: MPL-2.0

        mcpmeasured growthOpen source ↗

        Install claude mcp add llmtrim -- npx @llmtrim/cli

        222GitHub stars+6 (+2.8 %)
      • 10

        Skill
        Active

        Governance framework for AI coding agents. It runs them through a five-step workflow (plan, build, review, test, ship) where no step counts as done without evidence. Drop-in rules and guardrails for Claude Code, Codex, Cursor, Copilot, and…

        PythonLicense: MIT
        • agent-framework
        • agent-orchestration
        • agentic-development
        • agents-md
        • ai-agent

        githubmeasured growthOpen source ↗

        Install git clone https://github.com/KbWen/agentic-os ~/.claude/skills/agentic-os

        150GitHub stars+5 (+3.4 %)
      • 11

        Skill
        Active

        Lossless context compression plugin for Claude Code & OpenCode.

        TypeScriptLicense: BSD-3-Clause
        • anthropic
        • chatgpt
        • claude
        • claude-code
        • claude-code-plugin

        githubmeasured growthOpen source ↗

        Install /plugin marketplace add aerovato/magic-compact

        144GitHub stars+1 (+0.70 %)
      • 12

        Skill
        Active

        Content-aware output compression for AI coding assistants. 36 specialized processors cut CLI output tokens by 60-99% (git, pytest, npm, terraform, kubectl, docker, and more) without losing errors, diffs, or stack traces.

        PythonLicense: Apache-2.0
        • agentic-ai
        • ai-coding-assistant
        • claude-code
        • claude-code-plugin
        • cli-tools

        githubmeasured growthOpen source ↗

        Install /plugin marketplace add ppgranger/token-saver

        140GitHub stars+1 (+0.72 %)
      • 13

        Skill
        Active

        Claude Code usage governor: compact professional output, context slimming, tool-output filtering, telemetry, and drift guardrails.

        PythonLicense: MIT
        • ai-tools
        • claude-ai
        • claude-code
        • claude-code-plugin
        • claude-skills

        githubmeasured growthOpen source ↗

        Install /plugin marketplace add 0xhimanshu/governor

        133GitHub stars+4 (+3.1 %)
      • 14

        Skill
        Active

        Claude Code skill — delegate coding tasks to Mistral Vibe, supervise via git diff & save 50-90% Claude token usage

        Shell
        • claude
        • claude-ai
        • claude-ai-skills
        • claude-code
        • claude-skill

        githubmeasured growthOpen source ↗

        Install git clone https://github.com/pcx-wave/vibe-skill ~/.claude/skills/vibe-skill

        118GitHub stars+1 (+0.85 %)
      • 15

        MCP
        Active

        Tinder for your Claude Code skills, subagents and MCP servers. Swipe away what wastes context, scan for prompt injection, get honest token costs. Fixes what it finds and actually deletes. Zero dependencies, Claude Code + Codex.

        ShellLicense: MIT
        • agent-skills
        • ai-agents
        • claude
        • claude-code
        • claude-code-marketplace

        githubmeasured growthOpen source ↗

        115GitHub stars+1 (+0.88 %)
      • 16

        Skill
        Active

        Claude Code plugin that tracks token usage, identifies wasted context, and saves 30-50% on API costs. Heatmaps, ROI reports, budget alerts, efficiency scores, git-aware suggestions — all local, zero config.

        JavaScriptLicense: MIT
        • ai-tools
        • anthropic
        • claude-code
        • claude-code-plugin
        • cli-plugin

        githubmeasured growthOpen source ↗

        Install /plugin marketplace add egorfedorov/claude-context-optimizer

        102GitHub stars+7 (+7.4 %)
      • 17

        Skill
        Dormant

        Pith is the hook that makes Claude Code sessions last 3x longer.

        PythonLicense: MIT
        • anthropic
        • claude-code
        • claude-code-plugin
        • llm-tools
        • token-optimization

        githubmeasured growthOpen source ↗

        Install /plugin marketplace add abhisekjha/pith

        98GitHub starsstable
      • 18

        Skill
        Dormant

        A Claude Code skill for web report decks — richer ECharts charts, one file per slide (edit one page, save tokens), and complex structure diagrams. 16:9 & vertical, PDF/image export.网页版 PPT 汇报的 Claude Code skill：ECharts 把数据画得更好看、每页一个文件改一页省…

        HTMLLicense: MIT
        • ai
        • anthropic
        • claude
        • claude-code
        • claude-skills

        githubmeasured growthOpen source ↗

        Install git clone https://github.com/myunwang/ppt-report-skills ~/.claude/skills/ppt-report-skills

        81GitHub starsstable
      • 19

        Skill
        Active

        Agent Dashboard: Visualization and analytics for Sessions and Quota Usage. Track, analyze, and optimize token usage across providers with heatmaps, cost tracking, token counting and quota resets..

        PythonLicense: MIT
        • agent
        • analytics
        • anthropic
        • claude-code
        • claude-code-plugin

        githubmeasured growthOpen source ↗

        Install /plugin marketplace add JingbiaoMei/Tokdash

        66GitHub stars+3 (+4.8 %)
      • 20

        Skill
        Active

        SpiderBrain v3 is a multi-platform skill/framework to reduce token usage and AI hallucinations across Claude, Cursor, and other AI tools.

        • agent-skills
        • ai
        • ai-agents
        • anthropic
        • claude

        githubmeasured growthOpen source ↗

        Install git clone https://github.com/abhishek-performdigital/Spiderbrain-V3 ~/.claude/skills/Spiderbrain-V3

        65GitHub starsstable
      • 21

        MCP
        Active

        Context compression plugin for Claude Code. Automatically trims large tool output—JSON, YAML, stack traces, and logs—before it enters the context window.

        TypeScriptLicense: MIT
        • claude-code
        • claude-code-plugin
        • context-compression
        • context-window
        • developer-tooling

        githubmeasured growthOpen source ↗

        64GitHub starsstable
      • 22

        Library
        Active

        A high-performance, ergonomic web framework for Rust with native AI/LLM support.

        RustLicense: Apache-2.0
        • ai-native
        • api-framework
        • async
        • backend
        • developer-tools

        githubmeasured growthOpen source ↗

        61GitHub starsstable
      • 23

        Skill
        Active

        Own your LLM's web search: a local search->fetch->rank pipeline that replaces hosted web-search tools. Measured: matches hosted accuracy at 66% lower cost and up to 88% fewer tokens, plus a precision-tuned semantic caching with…

        PythonLicense: MIT
        • agent-memory
        • agent-skills
        • agentic-workflow
        • agents
        • anthropic

        githubmeasured growthOpen source ↗

        Install git clone https://github.com/firish/webfetch ~/.claude/skills/webfetch

        55GitHub stars+1 (+1.9 %)
      • 24

        MCP
        Active

        Cut context bloat in your AI-agent stack: find and safely prune unused skills, MCP servers and subagents from real transcript evidence

        GoLicense: MIT
        • agent-skills
        • ai-agents
        • anthropic
        • claude-code
        • claude-code-plugin

        githubmeasured growthOpen source ↗

        54GitHub stars+2 (+3.8 %)

      Learning resources

      Ranked by known GitHub stars, highest first; archived repositories come after maintained repositories. These resources remain available separately and do not take part in the main tool ranking.

      1. 1

        outsourcerer

        Skill
        ActiveOpen source ↗

        Make the most out of your subscriptions. Delegate work to other harnesses and models, while keeping your main session the orchestrator. Nothing new to learn. Keep working like you do, but better.

        Shell
        • agent-skills
        • agentic-workflow
        • claude-code
        • codex
        • fable
        github

        Install git clone https://github.com/alexgreensh/outsourcerer ~/.claude/skills/outsourcerer

        Resourcemeasured growth
        142GitHub stars+5 (+3.6 %)