EveryDev.ai
Subscribe
Home
Tools

5,901+ AI tools

  • Featured
  • Gems
  • Trending
  • New
  • Rate tools
  • Compare
  • Arena
Categories
  • Agents2972
  • Coding2248
  • Infrastructure1067
  • Projects818
  • MCP794
  • Research744
  • Analytics565
  • Marketing527
  • Design488
  • Testing476
  • Data442
  • Security412
  • Integration288
  • Communication284
  • Prompts283
  • Extensions238
  • Voice213
  • DevOps190
  • Commerce183
  • Web128
  • Learning123
  • Finance44
AI Tools by Topic
  • AI Coding Assistants
  • Agent Frameworks
  • MCP Servers
  • AI Prompt Tools
  • Vibe Coding Tools
  • AI Design Tools
  • AI Database Tools
  • AI Website Builders
  • AI Testing Tools
  • LLM Evaluations
Follow Us
  • X / Twitter
  • LinkedIn
  • Reddit
  • Discord
  • Threads
  • Bluesky
  • Mastodon
  • YouTube
  • GitHub
  • Instagram
Get Started
  • Users
  • Rate Tools
  • About
  • Editorial Standards
  • Corrections & Disclosures
  • Community Guidelines
  • Advertise
  • Contact Us
  • Newsletter
  • Submit a Tool
  • Start a Discussion
  • Write A Blog
  • Share A Build
  • Terms of Service
  • Privacy Policy
Explore with AI
  • ChatGPT
  • Gemini
  • Claude
  • Grok
  • Perplexity
Agent Experience
  • llms.txt
Theme
With AI, Everyone is a Dev. EveryDev.ai ©
    1. Home
    2. Tools
    3. Cache Tax
    Cache Tax icon

    Cache Tax

    LLM Extensions
    Featured

    A Claude Code mod that keeps the prompt cache warm during breaks and shows the estimated rewrite cost before a cold send.

    Visit Website

    At a Glance

    Pricing
    Open Source

    Free and open source Claude Code mod under the MIT License. Pings consume Claude tokens billed by Anthropic/the user's Claude plan, not by Cache Tax.

    Engagement

    Available On

    Windows
    API
    CLI

    Resources

    WebsiteDocsGitHubllms.txt

    Topics

    LLM ExtensionsContext EngineeringCompute Optimization

    Alternatives

    Maximem VitySoL-Piclaude-reflect
    Developer
    Karan BansalIndia

    Listed Nov 2026

    About Cache Tax

    Cache Tax is an open-source mod for Claude Code by Karan Bansal. It refreshes the session's prompt cache while you step away and asks for confirmation, with an estimated rewrite cost, before a cold prompt is sent. It requires Claude Code 2.1.287 or later and is distributed under the MIT license.

    What It Is

    Cache Tax is a Claude Code mod that works around the cost of an expired prompt cache. When the cached prefix expires, even a short message can trigger a rewrite of the whole old context. The mod addresses this in two ways: a bounded keepwarm window that pings the cache while you are away, and a guard that shows the estimated cost when you return to a cold session.

    How the Keepwarm and Guard Workflow Works

    Running /keepwarm 90m before a break arms a window. After 50 idle minutes, a timer inside Claude Code sends one tool-less fork over the session's transcript to refresh the cache. After each ping it reads usage and stops if reads are zero or writes reach 10% of reads. An optional idle spending limit (Ping floor) can skip later pings.

    For contexts of at least 50,000 tokens, the guard asks before an ordinary message is sent into a cache its one-hour clock says is cold. It shows the estimated rewrite cost and offers Send anyway or Don't send. Where nobody can answer, it refuses once and the resend goes through. By default, a detected cold write automatically arms keepwarm for at least three hours; the Auto keepwarm option in /config turns this off.

    Commands and Display

    • /keepwarm, /keepwarm 90m, /keepwarm always, /keepwarm status and /keepwarm off control warming.
    • /cache-tax shows the current cache estimate, warming window and the session's cold-write tally.
    • While keepwarm is armed, a row above the prompt (or a status entry) shows warm, cold or unknown state, the next ping and the last receipt.

    Costs, Limits and Reach

    Dollar figures are API-equivalent estimates, not extra subscription charges, and pings cost tokens, including uncapped output. The 50-minute schedule needs a one-hour main cache; API-key and some other billing paths default to five minutes and need promptCacheTtl set to "1h". The README documents the hooks and calls the mod uses and states that only the fork request leaves the machine. A separate hook-based version exists for people who prefer settings hooks.

    Cache Tax - 1

    Community Discussions

    Be the first to start a conversation about Cache Tax

    Share your experience with Cache Tax, ask questions, or help others learn from your insights.

    Pricing

    OPEN SOURCE

    Open Source

    Free and open source Claude Code mod under the MIT License. Pings consume Claude tokens billed by Anthropic/the user's Claude plan, not by Cache Tax.

    • MIT License
    • /keepwarm cache warming
    • Cold-send guard with estimated rewrite cost
    • Requires Claude Code 2.1.287+
    • Pings cost tokens (external Claude usage costs)

    Capabilities

    Key Features

    • Keepwarm window that pings the prompt cache after 50 idle minutes
    • Cold-send guard with estimated rewrite cost and Send anyway / Don't send choice
    • Automatic keepwarm arming after a detected cold write
    • Idle spending limit (Ping floor) for pings
    • Row above the prompt showing warm/cold state, next ping and last receipt
    • Per-session cold-write tally via /cache-tax
    • Per-model price rows for cost estimates
    • Headless prompt handling option
    • Read-only $.cacheTax.status() API for other mods

    Integrations

    Claude Code
    API Available
    View Docs

    Ratings & Reviews

    No ratings yet

    Be the first to rate Cache Tax and help others make informed decisions.

    Rate other tools you’ve used

    Developer

    Karan Bansal

    Head of AI at ArmorCode. Creator of Anya (autonomous ASPM agent). Building agentic AI for enterprise security. IIT Kanpur CS. 4500+ GitHub stars.

    India
    Read more about Karan Bansal
    WebsiteGitHubLinkedInX / Twitter
    1 tool in directory

    Similar Tools

    Maximem Vity icon

    Maximem Vity

    Cross-platform AI memory plugin that syncs context across OpenClaw, ChatGPT, Claude, and Gemini so you never re-explain yourself.

    SoL-Pi icon

    SoL-Pi

    A standalone extension for the Pi coding agent that adds four opt-in token-efficiency mechanisms found through scaled auto-research loops.

    claude-reflect icon

    claude-reflect

    A self-learning system for Claude Code that captures corrections and workflow patterns and turns them into CLAUDE.md memory and reusable skills.

    Browse all tools

    Related Topics

    LLM Extensions

    Extensions that enhance large language models with new capabilities.

    27 tools

    Context Engineering

    Techniques for optimizing context windows to improve AI responses.

    132 tools

    Compute Optimization

    Tools for optimizing computational resources and performance.

    93 tools
    Browse all topics
    Back to all toolsSuggest an edit
    ratings
    discussions