Hey, I'm Yevheny
I'm building Tessera — a programmable gateway that sits between your code and OpenAI / Anthropic / Mistral / Groq / Cohere, auto-routing requests to cheaper-equivalent models, caching the obvious repeats, compressing prompts losslessly, and arbitraging async traffic onto batch APIs.
You install the SDK, your existing code runs unchanged, and your LLM bill drops 30-40% on real workloads. Flat monthly subscription priced on gross token volume — you keep 100% of the savings.
Free Sandbox: 60M tokens/month, no card up front. Built for solo devs, side projects, hobbyists who want production-grade infra without a procurement conversation. Paid tiers: Starter $199 · Growth $999 · Scale $3,999 monthly. Enterprise custom above 20B.
- tessera-sdk — drop-in patcher for
openai,anthropic,mistralai,groq,cohereclients (Python + Node) - Framework adapters — LangChain, Vercel AI SDK, LlamaIndex, Mastra, Pydantic AI, CrewAI, AutoGen
- The proxy itself — 11 production mechanics with per-stack quality canary auto-rollback, audit-immutable pricing snapshots
Finance background before AI infrastructure. The audit-immutability discipline (every dollar number traces back to a pricing snapshot uuid) comes from that side of my career — cost work is more credible when the person doing the math has done it on someone else's money first.
Solo right now. Looking for a technical co-founder for the next phase. If you're interested in AI request infrastructure as a category — let's talk.
- Email: founder@tesseraai.io
- Site: tesseraai.io · tesseraai.io/dev · tesseraai.io/blog
Tessera is operated by Fintechagency OÜ, Tallinn, Estonia (registry 16638667).


