Credyt – Real-time billing, built for AI
Hey HN! I'm the founder of Credyt (https://credyt.ai). If you're running an AI product that burns $0.50+ per interaction, you need to bill customers as usage happens, not at the end of the month. Otherwise you're a bank for your customers, and you'll run out of money before they do. Why invoice-based billing doesn't work for AI: Most usage-based billing tools (Stripe Billing, Orb, etc.) were built for SaaS. They meter usage all month, then send an invoice. That works fine when your COGS is 5%. It's a disaster when you're paying your AI vendor upfront for every request and your…
What it does
In the maker’s words, at launch
Hey HN! I'm the founder of Credyt (https://credyt.ai). If you're running an AI product that burns $0.50+ per interaction, you need to bill customers as usage happens, not at the end of the month. Otherwise you're a bank for your customers, and you'll run out of money before they do. Why invoice-based billing doesn't work for AI: Most usage-based billing tools (Stripe Billing, Orb, etc.) were built for SaaS. They meter usage all month, then send an invoice. That works fine when your COGS is 5%. It's a disaster when you're paying your AI vendor upfront for every request and your margins are 15-30%. If a power user runs up $10K in costs and churns or their card declines at month-end, you've just eaten $10K. We've seen multiple startups nearly run out of runway because of this. "Just build a wallet in Postgres" – harder than it sounds: Everyone starts there. Track a balance, decrement on usage, done. Then you hit race conditions when usage spikes, auto-top-up logic that needs to be bulletproof, credit grants with expiry dates and priority rules, pricing versioning (customer X is on v1 rates, customer Y on v2, customer Z has a custom deal), profitability analysis (what did this customer cost us vs. what we charged them?), refunds, disputes, partial credits, reserves... Three months later you've hacked together a billing system instead of building your product. What Credyt does: It's a wallet-native billing engine. Your customers get a balance (in USD, credits, tokens, minutes, whatever you want), you send usage events as they happen, and Credyt handles the debiting, auto-top-ups via Stripe, and gives you an embeddable billing portal. We match against your pricing rules (which you can version and test before publishing), debit the wallet atomically, trigger top-ups via Stripe when they hit their threshold, and give you a hosted billing portal so you're not building balance UIs. Unlike tools that only track revenue, Credyt lets you log both what you charged and what it cost you. So you can see margins per customer, per product, per pricing tier—in real-time. Most founders are shocked when they realize their most active users are unprofitable. Docs: https://docs.credyt.ai Quickstart video: https://www.youtube.com/watch?v=48XJlId7098 Pricing: Free to start, then $1/month per active wallet. No processing markup—Stripe fees pass through directly. We're still early stage and genuinely want feedback from people in the trenches: What would stop you from ripping out your current system for this? What's missing that would make this production-ready for you? If you tried to build wallets yourself, where did it break?
Does the same job
all alternatives →
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, January 2026
the whole month →- IN
Hey HN! I wanted to share something I built over the last few weeks: isometric.nyc is a massive isometric pixel art map of NYC, built with nano banana and coding agents. I didn't write a single line of code. Of course no-code doesn't mean no-engineering. This project took a lot more manual labor than I'd hoped! I wrote a deep dive on the workflow and some thoughts about the future of AI coding and creativity: http://cannoneyed.com/projects/isometric-nyc
AI · Jan 2026 · cannoneyed.com
Automatic AI-powered code reviews the moment you open a PR
Dev tools · Jan 2026 · kilo.ai
