Topic

Tools for controlling AI token cost

217

we track

214

answered today

10

arrived this week

0

stopped answering

3 we couldn't reach today; we don't claim those are dead. Watching this group for 26 days, checked once a day · how we check

Codex Usage and Resets Plugin

LiveWatched 6d

Know your usage pace and resets. Let Codex help you plan.

FreeOpen source

What to know▼

How it works: İşlemi ayırır (detach), konuşmayı serbest bırakır, sınırlı çıktı kaydeder ve tamamlanınca Codex'e inceleme için döner.

CodexUsageBar

LiveWatched 13d

A free, open-source macOS menu bar app that tracks your ChatGPT Codex usage limits in real time for developers.

FreeOpen source

What to know▼

How it works: Lives in the macOS menu bar and shows current Codex usage at a glance.
Pricing: Free and open-source.

BlazePhoenix On-chain DEX Aggregator, Staking

LiveWatched 9d

Decentralized on-chain DEX quote aggregator and staking token across several blockchain networks.

What to know▼

How it works: On-chain DEX aggregator for token swaps and a staking engine for its own BZPX token, across Base, Ethereum, Optimism, and Arbitrum.
Pricing: Free.

ReserveGauge

LiveWatched 20d

A macOS menu-bar app that watches your AI coding usage on Codex or Claude Code and forecasts whether you'll hit your limit before it resets, all processed locally.

What to know▼

How it works: A macOS menu-bar app that forecasts your remaining usage for AI coding tools like Codex and Claude Code and flags whether your current pace will last until the next limit reset.
What's different: Processes usage data locally on-device rather than sending it to a server.
Best for: Developers on usage-limited AI coding plans who want to avoid unexpectedly hitting their limit.

AI Usage Ball

LiveWatched 6d

Live desktop orbs for Claude, Codex & Antigravity

FreeWorks offlineNo subscription

What to know▼

How it works: Kalan kullanım hakkı ve sıfırlanma zamanını yerelde izler, yalnız sağlayıcıların kendi servisleriyle konuşur; kaynak denetime açık.
Pricing: İlk 100 kullanıcıya ücretsiz ömür boyu kişisel lisans fırsatı.

LLMSlim

LiveWatched 17d

An open-source Python library that compresses LLM prompts and context to cut token costs while preserving instructions.

Open source

What to know▼

How it works: Compresses prompts, RAG document contexts, and multi-turn chat logs in one line of code.
Pricing: Open-source Python package.
Best for: Developers who want to cut LLM token costs 40-70% without losing instruction fidelity.

MCP-Recall

LiveWatched 9d

Compresses MCP tool outputs and stores full results in SQLite so agent sessions fit more tool calls.

What to know▼

How it works: Compresses MCP tool outputs (claims 94 KB to 3.5 KB, 96%) and stores full results in SQLite for retrieval, enabling up to 30x more tool calls per session.

Spanlens

LiveWatched 6d

Open-source, self-hostable observability platform for LLM apps: request logging, cost tracking, agent tracing.

Open sourceSelf-hosted

What to know▼

How it works: Works with OpenAI, Anthropic, and Gemini apps.
Pricing: Open source, MIT license.

SolReclaim

LiveWatched 15d

Recover your SOL account rental fees with one click.

Watch out · Includes a referral program paying 60% of the service fee in SOL — a financial incentive structure worth noting.

What to know▼

How it works: Closes unused Solana token accounts and recovers the locked SOL rent deposit (about 0.002 SOL per account) held in each one.
Best for: Solana wallet holders with leftover unused token accounts holding small locked rent deposits.
Watch out: Includes a referral program paying 60% of the service fee in SOL — a financial incentive structure worth noting.

Pylva

LiveWatched 12d

Tracks the real infrastructure cost of each customer for AI agent builders, so they can set margin caps per customer.

Open source

What to know▼

How it works: SDK-only telemetry tracks LLM and non-LLM usage per customer
What's different: open source, reactive rules like per-customer cost caps
Pricing: open source
Best for: builders of AI agent products billing by usage

Codex Quota Panel

LiveWatched 10d

A free open-source Windows dashboard that shows your Codex usage quota, reset timers, and token counts at a glance.

FreeOpen sourceNo subscription

What to know▼

How it works: Displays weekly quota, reset time, reset-credit cards, and lifetime token usage for Codex
Pricing: Free and open-source

Shibaita.ai

LiveWatched 7d

Lets Claude Code users compare usage on a public leaderboard instead of only seeing their own local stats.

Watch out · unofficial, not affiliated with Anthropic

Free

What to know▼

How it works: A CLI reads local logs and submits only aggregate token counts, never prompts or code.
Pricing: free
Watch out: unofficial, not affiliated with Anthropic

KlaatAI

LiveWatched 9d

Routes prompts to the cheapest capable model tier and offers a terminal coding agent claiming lower cost than Claude Code on a benchmark.

Free

What to know▼

How it works: Routes every prompt to the cheapest model tier that can handle it, escalating only when needed; its terminal agent Klaat Code claims to solve the same 33-task benchmark as Claude Code at 5.4x lower cost ($0.027 vs $0.146/task).
Pricing: Free beta: 75 requests/day, no card; tool calls are free, only messages count against quota.

LoRA Speedrun

LiveWatched 12d

Leaderboard site where people compete to fine-tune the same model fastest, with every record re-verified by a referee.

Free

What to know▼

How it works: same model, same score, same GPU — fastest training run wins, re-run 3x by a referee
What's different: no self-reported times, all records independently verified
Pricing: free to attempt

Local AI Art: PocketDiffusion

LiveWatched 13d

An offline iPhone app that generates AI images locally using Stable Diffusion, for people who want image generation without sending prompts to the cloud.

Works offline

What to know▼

How it works: Runs Stable Diffusion locally on iPhone, fully offline with no internet connection or API costs.
Best for: iPhone users who want AI image generation without sending data to the cloud.

CC Theme

LiveWatched 11d

An open-source Mac app providing ready-made, reproducible themes for AI desktop apps without repeated prompt generation.

Open source

What to know▼

How it works: Applies validated, ready-made theme packages to supported AI desktop apps, avoiding repeated image-and-prompt generation.

nable

LiveWatched 6d

One usage meter for every AI provider you pay for

FreeNo sign-upWorks offline

What to know▼

How it works: Kullanım yerel oturum kayıtlarından okunur, hesap/anahtar gerekmez; sabit planlarda kullanım, ölçülü planlarda tahmini dolar gösterir.
Pricing: Açık kaynak ve sonsuza dek ücretsiz.

ProxAI

LiveWatched 9d

Reverse proxy that enforces hard spending caps on LLM API calls across providers by swapping the base URL.

Watch out · AGPL-3.0, no telemetry.

FreeNo trackingSelf-hosted

What to know▼

How it works: A transparent reverse proxy enforcing hard hourly/daily spending caps per LLM API client via a one-line base_url change, returning 429 once the cap is hit rather than calling the upstream.
What's different: Works with OpenAI, Anthropic, Mistral, Groq, DeepSeek, and 5 more providers.
Pricing: Hosted beta at proxai.eu is free, no credit card; self-hostable as a ~10 MB Docker image.
Watch out: AGPL-3.0, no telemetry.

Toolport

LiveWatched 26d

A local gateway that lets every AI coding client share one set of MCP server connections instead of configuring each tool separately, cutting how many tools your agent has to load.

Watch out · The '91% overhead cut' figure is the product's own claim; we haven't independently benchmarked it.

FreeNo sign-upWorks offline

What to know▼

How it works: A local-first, open-source gateway. You set up and authenticate each MCP server once, and every connected AI client — Claude, Cursor, VS Code, Codex — shares that same connection instead of you configuring it per tool.
What's different: Claims up to 91% less tool-token overhead by giving an agent a handful of meta-tools instead of loading hundreds of individual tool definitions; API keys stay in the OS keychain.
Pricing: Free, open source, no sign-up.
Best for: Developers running multiple AI coding clients against the same set of MCP servers who don't want to re-authenticate each one separately.
Watch out: The '91% overhead cut' figure is the product's own claim; we haven't independently benchmarked it.

Edgee Codex Compressor V2

LiveWatched 10d

Benchmark and gateway that compresses context to cut Codex token usage and session cost.

What to know▼

How it works: A context-compression gateway for Codex that reduces token usage and boosts cache hit rate.
What's different: Benchmarked at 49.5% fewer input tokens, cache hit rate up from 76.1% to 85.4%, and 35.6% lower total session cost.

Oriveo

LiveWatched 14d

A synced multi-model AI chat app across iOS, Android, and web using your own API keys.

FreeWorks offline

What to know▼

How it works: Lets you chat with ChatGPT, Claude, Gemini, and 10+ other AI models in one app across iOS, Android, and web, synced in real time, using your own API keys.
What's different: Includes real-time cost tracking, local-first storage, saved Notes, and cross-checking replies with a second model.
Pricing: Try free, no key needed; no markup or extra subscription with your own keys.

Tokenez

LiveWatched 6d

Counts your AI usage locally across ChatGPT, Claude, Gemini, Claude Code and Cursor in one dashboard, showing your real total and how close you are to each limit.

FreeNo sign-up

What to know▼

What's different: Counts locally rather than relying on each provider's own scattered usage view.
Pricing: Free, no signup to try.

Token Guard

LiveWatched 8d

Tracks and converts LLM API spend across OpenAI, Anthropic and Gemini locally, instead of trusting a cloud dashboard with your usage data.

FreeWorks offlineOpen source

What to know▼

Pricing: Free and open source.
Best for: Developers juggling multiple LLM provider SDKs and costs.

Alenia Porter

LiveWatched 15d

A high-performance batch image/video/audio optimizer with GPU acceleration, for developers and creators needing fast local media compression instead of a cloud transcoding service.

What to know▼

How it works: A universal media optimizer that compresses images, video, and audio, built with Nuitka and Go, with auto-GPU acceleration and a crash-safe fallback.
What's different: Runs with GPU acceleration and a crash-safe fallback instead of relying on a cloud transcoding service.
Best for: Developers and creators who need fast batch compression of images, video, and audio.

Levitan

LiveWatched 11d

A free, fully offline voice-to-text tool that types what you say into any app on Windows or Mac, running locally with no cloud upload.

FreeWorks offline

What to know▼

How it works: hold a key, speak, text appears at cursor in any app; runs locally on CPU
Pricing: free forever
Best for: privacy-conscious users who don't want cloud-based voice typing

TripoSplat WebGPU on Vercel!

LiveWatched 16d

Generates a 3D model from a single image entirely on-device using WebGPU, with no server uploads or accounts needed — an open-source tool for developers experimenting with local image-to-3D generation.

Works offlineData stays with youOpen source

What to know▼

How it works: Generates a 3D model from a single image entirely on-device using WebGPU compute in the browser, with no server uploads or accounts.
What's different: Open source and runs locally via WebGPU rather than requiring cloud GPU processing, also deployable on Vercel for instant testing.
Best for: Developers experimenting with local, private image-to-3D generation.

Claude Usage Tracker

LiveWatched 10d

A Chrome extension showing your Claude usage limits, reset timers, and a projected time-left-at-this-pace right in the chat interface.

No sign-upWorks offline

What to know▼

How it works: tracks session/weekly limits locally and projects remaining time at current usage rate
What's different: fully on-device, no account needed
Pricing: free

MutualGPU

LiveWatched 10d

Lets people with spare GPU capacity share it so others without enough local compute can run demanding AI workloads.

Free

What to know▼

How it works: community compute-sharing network; first use case runs 3D Gaussian Splat generation via WebGPU
Pricing: free

Velokey

LiveWatched 14d

A pay-per-token API gateway giving developers access to 100+ AI models without a subscription commitment.

No subscription

What to know▼

How it works: Pay-per-token API gateway giving access to 100+ AI models.
Pricing: Pay per token, no subscription.
Best for: Developers.

Avery

LiveWatched 17d

Describes an agent in plain English and it builds a deterministic version that runs locally on your own hardware instead of a metered cloud API, for privacy-conscious teams that can't send sensitive work to the cloud.

Works offline

What to know▼

How it works: You describe an agent in one plain sentence, and Avery builds a deterministic version of it that runs on your own hardware rather than a metered public cloud, with no code required.
What's different: Runs locally and deterministically instead of routing through a per-token cloud API, aimed at avoiding cost, privacy, and lock-in problems of typical agentic AI projects.
Best for: Teams that want private, auditable automation they own outright, without sending sensitive work to a cloud API.

AI Usage Tracker

LiveWatched 12d

Local-first dashboard that auto-discovers your Claude and Codex spend across 10+ developer tools, with charts and zero telemetry.

Works offlineData stays with youNo tracking

What to know▼

How it works: auto-discovers usage across 10+ dev tools locally
What's different: zero telemetry, data stays on your machine
Best for: developers tracking AI tool spend

Meterbility

LiveWatched 10d

A debugger for AI agent runs that shows exactly what context each model call saw, for developers troubleshooting agents built with Claude Code, Codex, or Cursor.

Works offlineOpen source

What to know▼

How it works: Captures every step of an agent run, compares two runs, and can fork from any step to test changes
What's different: Local-first and open source, so traces never leave your machine
Best for: Developers debugging or comparing AI agent runs

Kadena Blockchain Development Company

LiveWatched 19d

A blockchain development agency building smart contracts, enterprise dApps, and token infrastructure.

What to know▼

How it works: An agency building custom blockchain products — smart contracts, enterprise dApps, token infrastructure and Web3 integrations — tailored to a client's commercial objectives.
Best for: Startups and enterprises wanting a development team to build production-ready blockchain products.

Codex Limits

LiveWatched 12d

A Mac menu-bar app that tracks your Codex usage limits, token activity and reset times across multiple accounts.

What to know▼

How it works: Native macOS menu-bar app and widget showing Codex rate limits, credits and reset times

UltraWork

LiveWatched 12d

Hosted coding environment with flat monthly pricing instead of pay-per-token billing for AI-assisted coding.

What to know▼

How it works: Flat-rate hosted environment with curated models and API-key access
What's different: Predictable flat pricing instead of per-token billing
Best for: Developers who want to avoid unpredictable AI token costs

Floe spend controls for voice AI agents

LiveWatched 12d

Gives voice-AI developers one account to meter spend across their whole vendor stack (LLM, speech-to-text, text-to-speech, telephony) with spend caps and kill-switches.

Open source

What to know▼

How it works: Server-side spend caps and per-agent kill-switches across multiple vendor APIs under one key
What's different: Open-source floe-guard component
Best for: Developers building voice AI products across several paid vendor APIs

ApiCredits

LiveWatched 9d

A prepaid AI API gateway giving one key to access multiple LLM providers at a lower cost than direct billing or aggregators.

No subscription

What to know▼

How it works: A single key and endpoint gives access to models from OpenAI, Anthropic, Google, xAI, DeepSeek, and Z.ai via OpenAI-, Anthropic-, and Gemini-compatible APIs, with built-in route health, load distribution, and fallback.
What's different: Priced lower than direct provider billing or aggregators like OpenRouter.
Pricing: Prepaid credits starting with a $5 test, no subscription or expiry.

OSGARD NEW WORLD

LiveWatched 9d

A tool that generates a themed sci-fi/fantasy AI universe with tradable in-game artifacts and a leaderboard.

FreeNo sign-up

What to know▼

How it works: Generates a themed AI universe (sci-fi, fantasy, cyberpunk or steampunk) with tradable artifacts in about 7 seconds, with a live marketplace, TC token staking and an AI assistant called JARVIS.
Pricing: 3 free generations, no signup needed.
Best for: People wanting to generate and trade AI-created world assets.

GPUCPUCompare

LiveWatched 9d

A free tool that compares GPU/CPU hardware specs and checks compatibility against a large game library.

FreeNo sign-up

What to know▼

How it works: Compares 733 GPUs and 1,002 CPUs and checks compatibility against 112,113 PC games using public TechPowerUp, Geekbench and Steam data.
What's different: Uses public benchmark data instead of unexplained percentages or simulated benchmarks.
Pricing: Free, no signup.
Best for: PC gamers deciding what hardware to buy.

PandaCoder

LiveWatched 9d

Free local AI coding assistant positioned as an alternative to token-limited cloud coding tools.

Free

What to know▼

How it works: Runs locally as a coding assistant with no token limits.
Pricing: Free.
Best for: Developers wanting a private, local alternative to token-limited cloud coding tools.

Dex Screener Clone Script

LiveWatched 9d

Ready-made clone script/template for building a DEX token-tracking platform.

What to know▼

How it works: A ready-to-develop template for building a platform that tracks token prices, trading pairs, liquidity, and volume across multiple DEXs.
Best for: Businesses wanting to launch a DEX tracking platform faster.

HygeiaCloud

LiveWatched 9d

Resells certified decommissioned enterprise GPUs (A100/H100/H200) at a discount to AWS/GCP pricing.

What to know▼

How it works: Intercepts decommissioned enterprise GPUs (A100/H100/H200) before they're shredded, certifies each unit through a 72-hour burn-in, and redeploys them.
What's different: 40-70% below AWS and GCP pricing, with zero new hardware manufactured.
Best for: Buyers needing serious GPU compute without new-hardware cost.

Find AI Credits

LiveWatched 9d

Free directory aggregating AI, cloud, API and GPU credit offers and startup perks.

Free

What to know▼

How it works: Free directory aggregating free AI, cloud, API, and GPU credit offers and startup perks from AI companies in one place.
Pricing: Free.

ChorusGraph

LiveWatched 23d

An open-source Python agent runtime, installed with one pip command, with a local cache that skips repeat LLM calls — pitched as a lighter alternative to LangGraph for developers.

Open source

What to know▼

How it works: An open-source agent graph runtime with a BSP scheduler and a local semantic cache that skips redundant LLM calls when a similar routing request already happened, installed with a single pip command.
What's different: Positioned as a lighter, non-wrapper alternative to LangGraph, with a local cache to cut LLM call costs.
Pricing: Open source (Apache-2.0).
Best for: Developers building LLM agent pipelines who want to reduce redundant model calls.

NagpurAdsWala

LiveWatched 10d

A performance marketing agency in Nagpur offering Meta Ads, Google Ads, SEO, and lead generation services.

What to know▼

How it works: A performance marketing agency in Nagpur offering Meta Ads, Google Ads, SEO, landing page optimization, and lead generation.

What Is a Million Tokens

LiveWatched 10d

Visually shows what a million AI tokens actually amounts to in books, code, or dollars, and compares model pricing.

Free

What to know▼

How it works: Converts token counts to relatable units and compares input/output pricing across models
Pricing: Free

Streaklee

LiveWatched 10d

A macOS menu bar app that shows how much of your paid Claude/ChatGPT usage quota you've used before it resets.

Works offline

What to know▼

How it works: Displays live usage windows, weekly capacity and reset times for AI subscriptions
What's different: Local-first, runs from the menu bar
Best for: People paying for Claude/ChatGPT/Codex who want to use their quota before it resets

Touki Blockchain

LiveWatched 10d

A no-code platform to create, deploy, verify, and use your own blockchain token.

What to know▼

How it works: Lets you create, deploy, verify, and use your own token on the Touki blockchain without needing development skills.
Best for: People who want to launch a token without being a developer.

TripoSplatWebGPU Local or Free MutualGPU

LiveWatched 10d

Generates 3D Gaussian Splats in your browser, for free, using either your device's GPU or a community-shared compute network.

FreeNothing to install

What to know▼

How it works: runs via WebGPU locally, or offloads to MutualGPU's crowd-sourced GPU network
What's different: no install, and free generation even without a powerful local GPU
Pricing: free

RamShared

LiveWatched 10d

Lets Linux/WSL2 users use idle GPU VRAM as extra swap memory without freezing the OS.

Watch out · Linux and WSL2 only

What to know▼

How it works: Cascading tier of zram, GPU VRAM, then NVMe with active latency monitoring and page demotion
What's different: Crash-safe unlike naive GPU swap tools, built with Rust/ublk/NBD drivers
Watch out: Linux and WSL2 only

AI Karma Tracker

LiveWatched 14d

A browser toolbar popup that shows live usage limits and reset times across the AI tools you use, with peak-hour alerts.

Watch out · No API keys — nothing leaves your browser.

Free

What to know▼

How it works: A toolbar popup showing live usage, remaining limits, and reset times for the AI tools you're signed into, with peak-hour nudges.
Pricing: Free for one tool, $2/mo for your whole stack.
Watch out: No API keys — nothing leaves your browser.

Ceiling

LiveWatched 14d

A Windows taskbar utility that tracks remaining usage limits and reset times across Claude, Codex, Cursor, and other AI tools.

Watch out · Windows only; works with Claude, Codex, Cursor, Gemini, and Copilot.

FreeWorks offline

What to know▼

How it works: Keeps your remaining AI capacity and reset times visible above the Windows taskbar, alerting you when a limit comes back early or partially restores; hovering shows per-model and per-project breakdowns and an equivalent API cost.
Pricing: Free and local-first.
Watch out: Windows only; works with Claude, Codex, Cursor, Gemini, and Copilot.

Smeltcore

LiveWatched 15d

Benchmark database showing which open-weight AI models run on which consumer GPUs, with install recipes and a GPU advisor.

FreeNo sign-up

What to know▼

How it works: Provides honest benchmarks for 84 open-weight AI models across 27 consumer GPUs from RTX 3060 to Apple Silicon, each pair rated as runs, tight fit, or won't fit with real speed and VRAM numbers, plus 800+ step-by-step install recipes for llama.cpp, Ollama, ComfyUI, and MLX, and a GPU Advisor.
What's different: No invented numbers.
Pricing: Free under CC BY-SA, no signup.

TTBox — Privacy,Safe,Quick Clipboard

LiveWatched 15d

An encrypted clipboard manager for developers to stash and quickly paste SSH keys, API tokens and server IPs from the menu bar, instead of keeping them in plaintext notes.

What to know▼

How it works: A menu-bar clipboard manager that stores SSH keys, API tokens and server IPs, encrypted locally with Argon2id and XChaCha20-Poly1305, with one-click or Enter-to-copy and sync across Mac, iOS, Windows and Web.
What's different: Zero-knowledge encryption applied specifically to developer secrets rather than general clipboard history.
Best for: Developers who need to quickly paste SSH keys, API tokens or server IPs without keeping them in plaintext notes.

AIVory Smart Inference

LiveWatched 16d

A hosted LLM inference router that's a drop-in replacement for the OpenAI API, automatically sending each request to the cheapest healthy provider. Aimed at developers who want lower inference bills without code changes.

No subscription

What to know▼

How it works: An LLM inference router and GPU marketplace that acts as a drop-in replacement for the OpenAI API, automatically routing each request in real time to the cheapest healthy provider serving the requested model.
What's different: Continuous automated price discovery across providers, rather than a fixed rate you sign up for once.
Pricing: No subscription.
Best for: Developers who want lower LLM inference bills without changing their SDK or code.

The Coin Lab

LiveWatched 20d

A no-code platform to launch your own crypto token across 12 blockchains in about a minute, for one payment, built for non-developers rather than blockchain engineers.

Pay once

What to know▼

How it works: A no-code platform to deploy ERC-20, SPL and BEP-20 crypto tokens live on 12 blockchains in about 60 seconds, without writing Solidity or using a CLI.
What's different: No developer or coding knowledge required to launch a token.
Pricing: One-time payment.
Best for: Non-developers who want to launch their own crypto token quickly.

RAI

LiveWatched 20d

A self-hostable AI chat platform that routes queries to the cheapest LLM and adds live market data, positioned as an alternative to ChatGPT or Claude you run yourself. For developers and cost-conscious teams.

FreeOpen sourceSelf-hosted

What to know▼

How it works: A self-hostable, open-source AI chat platform that runs on your own infrastructure, integrates live financial market data into chat, and routes each query to the best/cheapest LLM automatically across OpenAI, Anthropic, SiliconFlow and others.
What's different: Self-hosted with full data ownership and no vendor lock-in, plus intelligent model routing claimed to cut API costs by up to 60% versus using a single provider.
Pricing: Free forever to self-host, with Free / Pro / MAX tiers and per-model quotas.
Best for: Developers and cost-conscious teams who want a ChatGPT/Claude alternative running on their own infrastructure.

Live Crypto Ticker Basic and Premium

LiveWatched 16d

A dashboard for tracking live crypto prices and searching tokens across chains, with no account or subscription required.

Works offline

What to know▼

How it works: The Basic tier tracks Bitcoin, Solana, and top crypto assets via CoinGecko data on a split-flap terminal dashboard that runs locally; Premium adds DEX Screener search filtered by platform, origin, and blockchain.
Pricing: No subscriptions or setup required for the basic tier.

agentfdr

LiveWatched 22d

A debugging tool that reads Claude Code's own transcripts and turns them into a flight-recorder view showing loops, errors, and token burn, for developers using coding agents, instead of manually digging through raw logs.

Works offline

What to know▼

How it works: Reads the transcripts Claude Code writes and turns them into a per-turn timeline with anomaly flags (loops, error streaks, token burns), live watch mode, and cost estimates.
Pricing: Free and open source (MIT license), runs 100% locally with zero instrumentation.

FastDailyTools

LiveWatched 11d

A growing collection of free, no-signup online calculators covering AI tokens, fees, solar savings, and health tracking.

FreeNo sign-up

What to know▼

How it works: A collection of free calculators covering AI token estimation, PayPal/Etsy fees, solar savings, water intake, pregnancy due dates, and period tracking.
Pricing: Free, no sign-up or downloads required.

The ones we've read

A person-shaped pass went through these: someone read the maker's own description and wrote down how it works, what it costs and what to watch for.

Codex Analytics Ledger

LiveWatched 14d

A local-only dashboard for tracking Codex usage and cost, without sending usage data to the cloud.

What to know▼

How it works: A local-only dashboard for tracking Codex usage and cost.
What's different: Doesn't send usage data to the cloud.

Token Poker

LiveWatched 10d

Sprint planning poker tool that adds AI-token-usage estimates alongside story points for teams building AI features.

What to know▼

How it works: Teams vote on story points and forecast expected AI token consumption in the same planning session
What's different: Targets teams building LLM-based features where token cost affects planning
Best for: Software teams building AI-powered products who need to plan both effort and AI cost

Tokens Forge

LiveWatched 13d

An OpenAI-compatible AI gateway with transparent per-model pricing and usage tracking, paired with a separate AI tool for stock market research.

What to know▼

How it works: routes requests through a low-cost, OpenAI-compatible gateway with usage tracking, plus an AI stock researcher feature
Best for: developers wanting cheaper, trackable access to AI models

Clawgate

LiveWatched 18d

A cost-control tool that sets hard-stop budgets and tracks per-project spend for teams using Claude Code and Codex.

What to know▼

How it works: Applies budget limits and model controls to AI coding tool usage, with built-in prompt compression to cut token spend.
Best for: Teams that want visibility and caps on their Claude Code/Codex spend.

The new Firecrawl /search

LiveWatched 9d

Search API for AI agents that returns relevant excerpts instead of full pages to cut token usage.

Watch out · Live on every /search call today.

What to know▼

How it works: A search API that returns the excerpts from each result best matching the query instead of full pages.
What's different: Uses 10x fewer tokens than processing full pages and scores 94.7% on SimpleQA, the highest reported among providers.
Watch out: Live on every /search call today.

ZooData

LiveWatched 15d

Gives AI agents clean structured data (and pre-built e-commerce intel) instead of raw scraped HTML, for developers building agents.

What to know▼

How it works: Turns any URL into agent-ready structured JSON, and provides pre-analyzed e-commerce intelligence (competitor, market, traffic, consumer insights) live for Amazon and TikTok. API, CLI, and MCP server included.
What's different: Uses about 75% fewer LLM tokens than working with raw HTML or markdown, and charges only for the fields you use.
Pricing: Starts with 1,000 free credits, no card required.
Best for: Developers building AI agents that need clean structured web or e-commerce data.

agentrecap

LiveWatched 6d

Simple tool to analyze your local coding-agent sessions

What to know▼

How it works: Yerel coding-agent oturumlarını okuyup çalışma süresi, yapılan araç çağrıları, değiştirilen kod satırı ve mesaj sayısı gibi veriler çıkarır.

The Builder Index — Watch Me Pivot

LiveWatched 10d

Public leaderboard of developers' AI tool usage, tracking token spend and surfacing effective AI setups.

What to know▼

How it works: You connect Claude and Codex to a public board that tracks your token usage and surfaces the skills, tools, and projects you use.

Just arrived

Landed in this group in the last seven days. No track record yet, and we say so rather than hiding them.

Tempest

LiveFirst seen today

Track AI

LiveWatched 1d

Tokimeter

LiveWatched 1d

No sign-upOpen sourceNo tracking

97AI.PRO

LiveWatched 2d

Free

AI API Cost & Budget Tracker

LiveWatched 2d

TokenDam

LiveWatched 2d

FreeData stays with you

PackMD

LiveWatched 2d

SendPI

LiveWatched 2d

Watched the longest, still answering

The ones we have been checking the longest that still respond today. This is the whole point of the site: not what launched, what survived.

Auriko

LiveWatched 24d

Trading desk for LLM calls

Watch out · The '30% average cost reduction' figure comes from the product's own published report, not an independent benchmark.

What to know▼

How it works: Sits between your app and multiple LLM providers, automatically routing each request to the cheapest/fastest provider based on token price, cache behavior, latency and reliability.
What's different: Built by ex-quant traders, applies trading-desk arbitrage logic to LLM API pricing rather than simple static routing rules.
Best for: Teams making heavy, repeated LLM API calls who want automatic cost optimization without manually comparing providers.
Watch out: The '30% average cost reduction' figure comes from the product's own published report, not an independent benchmark.

Constellation Gate AI

LiveWatched 24d

A desktop proxy that sits in front of your AI agent's calls to block prompt injection, scan for leaked secrets and cut token usage - works alongside your existing Claude or ChatGPT subscription, not a replacement for it.

Watch out · The '#1 in 16 benchmarks' and '20-40% token savings' figures are the vendor's own claims.

What to know▼

How it works: A desktop app that proxies your AI agent's calls, adding prompt-injection defense, secret scanning, an audit trail, and prompt compression/caching. Works with your existing Claude or ChatGPT subscription or 100+ pay-as-you-go models.
What's different: Claims the #1 rank across 16 public prompt-injection benchmark datasets — a specific, citable security claim.
Best for: Developers running AI agents in production who want prompt-injection defense and token-cost savings without changing models.
Watch out: The '#1 in 16 benchmarks' and '20-40% token savings' figures are the vendor's own claims.

Dali by Lulu

LiveWatched 24d

An MCP tool that scores your AI image or video prompt before you generate, aiming to catch weak prompts before you spend a paid generation credit.

What to know▼

How it works: An MCP tool that scores an AI image or video generation prompt before you run it, aiming to flag weak prompts before a paid generation credit is spent.
What's different: Scores prompts pre-generation rather than after, and runs a live dashboard tracking community-wide savings from credits not wasted.
Best for: People generating AI images or video who pay per generation and want to avoid burning credits on bad prompts.

Make Claude Code cheaper and smarter

LiveWatched 24d

A lightweight proxy that sits in front of coding-agent tool calls to block repetitive failing calls, mutate strategy when the agent is stuck, and compress bloated tool outputs, working across Claude, OpenAI, DeepSeek, Ollama, and OpenCode.

What to know▼

How it works: Runs as a background proxy (installed via pip) that detects and blocks repetitive failing tool calls, mutates strategy on loops, and compresses tool outputs, with a real-time dashboard and savings report.
Best for: developers running coding agents who want to cut token waste and loops.

ManjuAi

LiveWatched 24d

Service that runs deep-research and coding agents using various AI models for a low flat fee.

Free

What to know▼

How it works: Gives access to deep-research and coding agents with control over report length and full context for coding tasks.
Pricing: Access to most premium models for $5.30.

SAGE - CLI

LiveWatched 24d

A command-line tool for developers that compresses terminal output to cut token usage and cost when working with AI coding assistants.

What to know▼

How it works: A command-line tool that compresses terminal output by 85-98% to reduce AI coding assistant token usage, keeping raw logs local and tracking real savings.
Best for: Developers using AI coding assistants who want to cut token costs from verbose terminal output.

AfterLimit

LiveWatched 24d

Queues a follow-up prompt and auto-sends it to ChatGPT, Claude or Gemini once your usage limit resets, so you don't lose your train of thought waiting around.

What to know▼

How it works: You type or paste a follow-up prompt, pick a send time (such as when your usage limit resets), and AfterLimit automatically sends it to ChatGPT, Claude, or Gemini at that time.
Best for: People who hit free usage limits on AI chat tools mid-task and don't want to remember to come back.

ReplyWatcher

LiveWatched 24d

Watches one person's public X posts for signs Codex quota has reset, then pushes ranked alerts (confirmed/possible/noise) to Discord or Lark.

Watch out · Relies entirely on monitoring one individual's public posts as the reset signal — if that account goes quiet or changes behavior, the alerts stop being reliable.

What to know▼

How it works: Watches one specific person's public X posts, uses AI semantic filtering to separate real Codex quota-reset signals from noise, then pushes ranked alerts (confirmed/possible/noise) to Discord or Lark.
Best for: Codex users who repeatedly hit usage limits and want to know the moment a reset window opens.
Watch out: Relies entirely on monitoring one individual's public posts as the reset signal — if that account goes quiet or changes behavior, the alerts stop being reliable.

Free

8 of 16

See the group →

PokeTokenBar

LiveWatched 11d

A menu-bar app that turns your Claude Code/Codex/Gemini CLI token usage into a Pokémon-collecting game while tracking usage limits.

FreeWorks offlineOpen source

What to know▼

How it works: Reads local usage logs and hatches/evolves a Pokémon as you spend tokens
What's different: Runs entirely on-device, nothing sent anywhere
Pricing: free
Best for: Developers who want a fun way to watch their AI token usage

TokenPulse

LiveWatched 12d

A free Chrome extension that shows a live token/context bar above Claude, ChatGPT, and Gemini input boxes so you don't get cut off mid-session.

FreeNo sign-up

What to know▼

How it works: Displays context window usage, rate limits, reset countdowns, and cost live above the input box
Pricing: Free, no account, no API key
Best for: Developers who use multiple AI chat tools daily and hit context limits

PXINK

LiveWatched 12d

A free browser tool that shrinks big text/code pastes into dense PNGs so AI vision billing costs less than text billing.

FreeData stays with you

What to know▼

How it works: Wraps text blocks into dense 1928px PNGs the model reads via OCR
What's different: Exploits vision-token vs text-token pricing gap for ~60% savings
Pricing: Free, client-side, no account
Best for: Developers pasting large text/code logs into Claude or Gemini

Palet

LiveWatched 23d

A free color palette generator and design-token tool for designers and developers, extracting colors from images and exporting to CSS, Tailwind, or design tokens.

FreeCan export my data

What to know▼

How it works: Generates palettes, extracts colors from images, checks contrast, and exports to CSS, Tailwind, and design tokens.
Pricing: Free.

Claude session monitor

LiveWatched 19d

A free local VS Code/Cursor extension that turns your Claude Code session files into a browsable dashboard with full transcripts, tool calls, sub-agents and per-response token and cost tracking.

FreeWorks offline

What to know▼

How it works: A VS Code/Cursor/VSCodium extension that turns local Claude Code session files into a browsable workspace with full transcripts, tool calls, sub-agents, per-response token and cost tracking, an analytics dashboard, and a CLAUDE.md memory view.
What's different: 100% local-first — nothing leaves your machine.
Pricing: Free.
Best for: Developers using Claude Code who want to review sessions, costs and tool calls without a CLI-only cost counter.

Packforai

LiveWatched 19d

Converts PDFs, Word, Excel, PowerPoint, CSV and JSON files into clean Markdown formatted for feeding into ChatGPT or Claude, free to start with no account required.

FreeNo sign-up

What to know▼

How it works: Converts PDFs, Word, Excel, PowerPoint, CSV, and JSON files into clean Markdown formatted to work well as input for AI models like ChatGPT and Claude.
What's different: Optimizes the output specifically to use fewer tokens and produce better AI answers, rather than just generic format conversion.
Pricing: Free to start, no account needed.
Best for: People who need to feed messy documents into ChatGPT or Claude and want cleaner, more token-efficient input.

Converly Colors

LiveWatched 19d

Generates AI color palettes as design.md token files for tools like Claude, Cursor, and v0 — a quick color-scale generator for developers and designers doing AI-assisted work.

FreeNo sign-up

What to know▼

How it works: Generates AI color palettes as design.md token files — the format used by tools like Claude, Cursor and v0 — including 12-step scales, light/dark modes and CSS variables.
What's different: Outputs directly in the design.md token format that AI coding tools expect, rather than a generic palette export.
Pricing: Free, no signup.
Best for: Developers and designers building AI-assisted workflows who need design tokens fast.

Honey (I Shrunk the AI) by GreenPT

LiveWatched 20d

Honey is a free, open-source (MIT) skill that reduces token usage for AI coding agents like Claude Code by enforcing concise code and prose conventions and cheaper bulk reads.

Free

What to know▼

How it works: enforces YAGNI/stdlib-first code, answer-first prose, compact JSON/ESON handoffs between agents, and renders bulk reads as images to cut token cost
Pricing: free, MIT license
Best for: developers running AI coding agents (Claude Code, GPT, etc.) who want lower token usage

Open source

8 of 11

See the group →

SuperCompress

LiveWatched 23d

Open-source API that compresses prompts before OpenAI, Claude, Gemini or RAG calls, cutting token costs by about 65% while keeping the important content. For developers wiring it into their own LLM calls.

Open source

What to know▼

How it works: An open-source API that compresses prompts before they're sent to OpenAI, Claude, Gemini, RAG, chatbot, or agent calls, stripping tokens while preserving the evidence needed to answer correctly.
What's different: Claims to remove more than 65% of tokens on average while preserving over 98% of answer-critical content.
Best for: Developers wiring their own LLM calls who want to cut token costs without integrating a full new pipeline.

StackSpine

LiveWatched 12d

An open-source control plane that enforces policy, budget limits, and compliance auditing across all the AI providers a company uses in production.

Open source

What to know▼

How it works: Sits between apps and AI providers to enforce guardrails and log usage for audit.
Pricing: Open-source, Apache 2.0.
Best for: Engineering teams running production AI across multiple providers who need governance.

burnban

LiveWatched 18d

A local, open-source binary that meters and caps AI agent spend on your own machine for developers who want budget control without sending usage data to a dashboard.

Open source

What to know▼

How it works: A local binary sits in the request path, meters every token spent by AI agents, and enforces spend caps before they're exceeded; also compares your flat-rate plan against real metered rates.
What's different: Open source (MIT); sync and team budgets are planned but not yet available.
Best for: Developers running AI agents who want local spend control without a hosted dashboard.

VOLY

LiveWatched 16d

Open-source control plane that routes coding tasks across AI agents like Claude Code, Cursor, and Codex with cost tracking.

Open source

What to know▼

How it works: An open-source control plane that routes coding tasks across AI coding agents (Claude Code, Cursor, Codex, Wrangler, etc.), runs multi-agent workflows, caches to reduce repeated token spend, falls back on billing/availability issues, and shows cost per task with files touched.
What's different: Gives one measurable layer above multiple separate agent tools instead of running each independently.

HardwareMon

LiveWatched 18d

Cross-platform hardware monitoring with real-time stats, history and benchmarking for Windows, Linux, macOS and Android, with all data processed locally and no telemetry upload.

Works offlineOpen source

What to know▼

How it works: Monitors CPU, GPU, memory, disk and network in real time, with historical analytics, process management, local benchmarking, and gaming session detection, across Windows, Linux, macOS (Apple Silicon), and Android.
What's different: All monitoring data is processed on-device with local SQLite storage; no telemetry or benchmark results are uploaded.
Best for: Users who want cross-platform hardware monitoring without sending data to the cloud.

CodexBar Lite

LiveWatched 7d

Lets developers monitor their OpenAI Codex usage from the macOS menu bar instead of checking a web dashboard, using only their existing CLI session.

Data stays with youOpen source

ISONGraph

LiveWatched 5d

Open source

TokenTelemetry

LiveWatched 3d

Works offlineOpen source

Works offline

4 of 8

See the group →

Prompt Compass

LiveWatched 12d

Routes each AI prompt to the cheapest capable model and blocks PII/jailbreak attempts locally, cutting AI API costs for developers.

Works offline

What to know▼

How it works: On-device ~5ms routing decision per prompt
What's different: Claims up to 96% cost cut by avoiding always-expensive-model routing
Pricing: Free tier, npm SDK and VS Code/Cursor extension

AgentQuartz

LiveWatched 4d

Claude & Cursor usage in your macOS menu bar

FreeWorks offline

Hydra

LiveWatched 4d

FreeWorks offline

LMspend

Watched 17d

A free, local-first CLI that reads logs already on your machine to estimate spend on AI coding tools like Claude Code and Cursor.

FreeWorks offline

What to know▼

How it works: A local-first CLI that reads logs already on the machine (no proxy, no traffic in the middle) to estimate spend on AI coding tools like Claude Code, Cursor, and Codex from published model pricing, printing monthly reports and shareable cards.
Pricing: Free; install via 'npx lmspend'. Optional dashboard adds history, budgets, alerts, and team roll-ups.
Best for: Individuals and teams wanting to track AI coding tool costs before the invoice arrives.

Tools for controlling AI token cost — the short answers

Every number here comes from our own daily check — not from a vendor list.

Tools for controlling AI token cost — how many are there?: Tablif is tracking 217 of them. 214 answered our check today, and 3 we couldn't reach — we don't claim those are dead.
Which ones are still maintained?: Tablif knocks on every door once a day and records the answer. 214 of these 217 responded on the latest run, so that number is what "still here" means on this page — not a review score.
Are any of them free?: 44 of the live ones say so in their own words, and 14 let you start without making an account. Tablif records the claim the product makes; we don't verify pricing.
Any open-source options?: 27 of the live ones on Tablif mention being open source.
What's the newest one?: Tempest — Tablif first saw it today.
Did a person actually look at these?: 215 of them Tablif opened and read, then wrote a one-line summary in our own words instead of reusing the founder's tagline. The rest carry keyword labels we haven't confirmed by reading yet — and we say so rather than hiding it.