TokenTally — AI cost clarity

Estimate how every request hits your budget before it lands in production.

🚀 What is TokenTally?

TokenTally is a Next.js app that centralizes pricing knowledge for today’s leading AI providers. It combines per-model calculators, live comparison tables, and visual benchmarks so teams can predict spend, plan deployments, and share numbers with stakeholders in seconds.

✨ Key Capabilities

• Core platform: Guided calculators for OpenAI, Anthropic, Google Gemini, Cohere, Perplexity.ai, Whisper/TTS, and DALL·E. Each flow supports per-model inputs, prompt caching details, and fine-tuning rates when available.
• Pricing intelligence: A global table driven by config/pricing.ts surfaces all models with search, provider/category filters, and last-updated metadata (PRICING_LAST_UPDATED). Teams can export full datasets from the header or inspect every entry in /models.
• Decision-ready visuals: Interactive cost and context window charts (components/pricing-charts.tsx) make it easy to compare providers side by side and justify model swaps with data.
• Fast workflows: A command palette (Cmd+K) accelerates navigation, while responsive layouts, dark mode, and animated cards keep the UI polished on any device.
• Feedback loop: Inline feedback routes (/feedback) and Sonner toasts encourage users to report outdated rates so pricing stays accurate.

🎯 Perfect For

• Product teams benchmarking LLM providers before shipping new AI features.
• FinOps and procurement partners who need repeatable pricing snapshots.
• Developer advocates preparing talks, workshops, or internal docs about AI cost trade-offs.

Quick Start

git clone https://github.com/gfargo/token-tally.git
cd token-tally
npm install
npm run dev
# open http://localhost:3000

git clone https://github.com/gfargo/token-tally.git
cd token-tally
npm install
npm run dev
# open http://localhost:3000

Technology Stack

• Next.js 15 App Router with client-side animations via Framer Motion.
• Tailwind CSS v4 utilities plus Shadcn UI and Radix primitives for accessible components.
• Recharts for cost/context visualizations.
• Sentry and Vercel Analytics instrumentation ready for production diagnostics.

🔗 Links

• Production: https://tokentalley.griffen.codes/
• GitHub: https://github.com/gfargo/token-tally
• Feedback form: https://tokentalley.griffen.codes/feedback

TokenTally keeps AI experiments grounded in real numbers—spin it up when you’re ready to forecast the next launch.