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 creation date, newest first; undated entries come last. 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

    Skill
    Active

    The Agent Knowledge Kit & Architecture System — 10–32x token-efficient architecture discovery for AI coding agents

    githubestimated momentumOpen source ↗

    Install git clone https://github.com/chama-x/quiv ~/.claude/skills/quiv

    3GitHub stars
  2. 2

    MCP
    Active

    Context optimization for LLM agents: tool registry, result masking, budgeting, compaction.

    mcpmeasured growthOpen source ↗

    Install claude mcp add distil -- npx @munhq/distil

    1GitHub starsstable
  3. 3
    Active

    A Claude Code hook that routes each Read on a PDF to the path that actually works: text layer to UTF-8 text, scans to vision, and neither one through poppler.

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add WatsonTsai/pdf-text-router

    7GitHub starsstable
  4. 4

    Skill
    Active

    Profile Claude Code context, discover repeated workflows, and regression-test your skills. Local-only, zero-dependency.

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add nks0614/harnessay

    5GitHub stars+2 (+66.7 %)
  5. 5

    Skill
    Active

    Cut Claude Code token costs by delegating bulk LLM work to cheap models. 600+ models via NanoGPT, OpenRouter, Groq or local Ollama — batched, resumable, with real per-call cost tracking.

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add bicced/sidequest

    1GitHub starsstable
  6. 6

    Agent
    Active

    Experimental no-install semantic language for AI agents: typed action/state, safe NL/JSON fallback, and falsifiable public evals. Try the one-file probe.

    githubestimated momentumOpen source ↗

    1GitHub stars
  7. 7
    Active

    Lossless context compression: 2-8x fewer tokens, byte-exact recovery, search inside payloads

    mcpmeasured growthOpen source ↗

    Install claude mcp add densely -- uvx densely

    6GitHub starsstable
  8. 8
    Active

    Claude Code skill: caveman-style token efficiency, spoken with a real Spanish regional accent

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add Mun1to/SpanishCaveMan

    2GitHub starsstable
  9. 9
    Active

    A skill that shrink noisy log/output 60-90% before read — filter, group, dedupe, truncate. Cut token, cut cost, keep signal. By Dhayalan (aka Joy).

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add Da-ya7/PromptCompressor-skill

    1GitHub starsstable
  10. 10
    Active

    Installable agentic skills (SKILL.md) for Claude Code, Cursor, Codex, Gemini & Antigravity — 292+ product, efficiency, and common-sense skills. npx major-ai-skills

    githubestimated momentumOpen source ↗

    Install git clone https://github.com/alivirgo/Major-AI-Skills ~/.claude/skills/Major-AI-Skills

    1GitHub stars
  11. 11
    Active

    Measure and shrink the token cost of Claude Code skills, tier by tier: the always-on system-prompt tax, the per-invocation body, and on-demand references, each priced at the rate it is actually billed.

    githubmeasured growthOpen source ↗

    1GitHub starsstable
  12. 12
    Active

    Reversible context compression for AI agents: cut token usage up to 94%, originals retrievable

    mcpmeasured growthOpen source ↗

    Install claude mcp add token_optimizer -- uvx slimctx

    1GitHub starsstable
  13. 13
    Active

    Statically audits MCP tool surfaces for token cost, schema quality, and design issues.

    mcpmeasured growthOpen source ↗

    Remote server, to be declared in the MCP configuration https://mcplint-web.vercel.app/api/mcp

    0GitHub starsstable
  14. 14
    Active

    Verifies token/cost savings claimed by AI context-reduction proxies via MCP tools.

    mcpmeasured growthOpen source ↗

    Install claude mcp add tokentrust -- uvx tokentrust-cli

    0GitHub starsstable
  15. 15
    Active

    Token cost math for LLM API calls: verified per-1M-token rates for 69 models, 17 providers.

    mcpmeasured growthOpen source ↗

    Install claude mcp add comparedge-llm-cost -- npx @comparedge/llm-cost-mcp

    2GitHub starsstable
  16. 16

    Skill
    Active

    Claude Code plugin: Fable 5 as a token-frugal orchestrator with tiered Opus/Sonnet/Haiku agents

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add realgarit/fable-baton

    22GitHub stars+1 (+4.8 %)
  17. 17

    Skill
    Active

    🪒 YAGNI enforcement at the harness level, nudges toward the simplest solution and gates against unnecessary dependencies, file sprawl, and unchecked build growth.

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add V-Songbird/razor

    8GitHub stars+1 (+14.3 %)
  18. 18

    Skill
    Active

    🤫 Token-lean sessions at the harness level. An easy to grasp output style, output-shrinking hooks, and log compression cut both input and output tokens.

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add V-Songbird/hush

    47GitHub stars+5 (+11.9 %)
  19. 19
    Active

    Lossless context compression plugin for Claude Code & OpenCode.

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add aerovato/magic-compact

    146GitHub stars+2 (+1.4 %)
  20. 20

    Agent
    Active

    Control plane for AI coding agents: route tasks, reduce token spend, run multi-agent workflows, fallback executors, and track cost per task.

    githubmeasured growthOpen source ↗

    16GitHub starsstable
  21. 21
    Active

    Eco mode for Claude Code. /eco: -31% to -73% output tokens with critical findings intact; /eco-max: up to -75% with lowered effort. Measured hardest on Claude Fable 5 (fable5), deep-studied on Sonnet 5, works on Opus 4.8 too. We publish our…

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add sup3x/claude-code-eco

    33GitHub starsstable
  22. 22

    Skill
    Active

    Tech-lead orchestration for Claude Code — the top-tier model (Fable) keeps architecture & decisions, cheap subagents (Sonnet/Opus) do the routine and the digging. Save tokens without losing quality.

    githubmeasured growthOpen source ↗

    Install /plugin marketplace add AndyShaman/senior-fable

    23GitHub starsstable
  23. 23
    Active

    Pay-per-call agent APIs over x402: web scraping, token compression, and semantic cache.

    mcpmeasured growthOpen source ↗

    Install claude mcp add gate402-mcp -- npx gate402-mcp

    2GitHub stars+1 (+100.0 %)

Learning resources

Ranked by creation date, newest first; undated entries come last. These resources remain available separately and do not take part in the main tool ranking.

  1. 1
    ActiveOpen source ↗

    Molavi Agent Skills: Curated AI agent skills and MCP configs for Antigravity, Cursor, Codex & Claude — by Taghi Molavi

    githubResourcemeasured growth
    8GitHub starsstable
  2. 2
    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.

    github

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

    Resourcemeasured growth
    142GitHub stars+4 (+2.9 %)