Shipped: TokenCost — AI Model Pricing Comparison
Second ThirkleBot MVP deployed. Compare 12 AI models side by side with real-time cost calculations. Cache pricing. Batch estimator. Free.
By JTT
Solo Developer & Writer
TokenCost is live at https://mvp-tokencost.vercel.app — the second product out of ThirkleBot’s autonomous pipeline. Built and deployed in under 3 hours.
What it does: Side-by-side pricing comparison for 12 AI models — GPT-4.1 (all sizes), Claude Opus/Sonnet/Haiku 4, DeepSeek V3/R1, Gemini 2.5 Pro/Flash, o3/o4-mini. Every price from official API docs. Toggle cached input pricing to see the real savings from prompt caching (50-75% lower). Batch cost estimator for document processing pipelines.
Why this exists: Hacker News had two massive threads this week. DeepSeek’s permanent price drop earned 544 points. Memory costs hitting two-thirds of AI chip costs earned 326 points. Everyone wants to understand what AI actually costs at scale. Nobody had built a dead-simple tool that shows the answer in 3 seconds.
The result: A tool that’s genuinely useful. Choose your input/output token counts and every model recalculates instantly. The cheapest option gets highlighted. Models sort by cost automatically. The batch calculator estimates monthly processing costs for 10,000+ documents through any model. All prices verified against official API docs.
Tech: Astro 6 + React 19 + Tailwind CSS. Single page. No database. No accounts. Plausible analytics. Dark mode. Same quality bar as Pricify — just aimed at a different market gap.
What changed: Build time dropped from 4 hours (Pricify) to under 3 hours (TokenCost). The pipeline is getting faster. The scaffold pattern works. Next build will be even quicker.
Get updates when I ship
Join the newsletter for build logs, tutorials, and product launches. No spam, unsubscribe anytime.
Written by Jordan Thirkle
Stay-at-home dad shipping AI-accelerated products. Posts come from real builds, not theory.
Continue Reading
Building Without Tech Debt: An AI Development Playbook
How to integrate AI into your development workflow without accumulating deprecated patterns, untested code, and unpayable technical debt.
Astro 6 View Transitions: High-Performance DOM Caching
Master Astro 6 view transitions. Learn how to architect zero-layout-shift DOM caching and unblock the event loop for highly performant, minimalist web apps.
Edge-Native Micro-SaaS with Astro 6 and Preact
A deep dive into building ultra-fast, edge-native micro-SaaS applications using Astro 6, Preact, and Cloudflare Workers. Zero-BS engineering for solo devs.