In the next few years, every digital action will be mediated by AI. Whoever owns your agent owns your context, your decisions, and your output.

I build and operate ODEI.AI.

ODEI.AI is becoming memory, governance, receipts, and local runtime access for agents working with humans. The human stays the owner. The agent becomes the operator. ODEI.AI is not another chatbot; it sits one layer above the AI tools (Claude, ChatGPT, Gemini, and others) that people already use, and gives those tools structured access to the human's working context — goals, rules, decisions, tasks, evidence, outcomes.

The product is delivered as a web application at app.odei.ai, supported by a public API at api.odei.ai. Both are operated by me as the sole proprietor. Service delivery is fully digital; no physical inventory is involved.

Modern AI tools are general-purpose. They do not know the human's actual work, decisions, history, or rules. The human spends significant time re-explaining context every session, every new chat, every new tool. ODEI.AI exists to close this gap.

The framing matters. The human stays the owner of their context, their data, and the rules under which the agent operates. The agent becomes the operator — it executes against memory the human has structured, with receipts the human can verify. This is what ODEI.AI is becoming.

The AI model is no longer the moat. Memory, governance, and receipts are. ODEI.AI does not compete with the AI labs (OpenAI, Anthropic, Google) that train the underlying models — it sits one layer above and gives the human what those models do not: a place for their data to live, be governed, and produce receipts.

Honest limitations. The product is in its early commercial stage. It does not replace enterprise CMS, project tools, or vertical AI products. The scope is narrow on purpose: memory, governance, receipts, and local runtime access for agents working with humans.

Revenue comes from four streams, all denominated in EUR or USD, all routed through Hungarian sole-proprietor accounts in accordance with local tax law. The interactive model below lets any reader edit assumptions and see projections update live.

Assumptions. Direct costs include hosting, database, AI model API spend, payment processing. Operating costs include software subscriptions, accounting, legal, occasional EU travel, insurance. The Hungarian sole-proprietor minimum income requirement is met from 2026 onward.

No external capital. The business operates without investor funding. Working capital comes from operations. No loans or debt obligations outstanding.

The addressable audience is large, growing, and underserved by current AI products. ODEI.AI does not need a large share to be sustainable as a sole-proprietor activity.

Capture target. Fewer than 300 active Pro subscribers (€19 / month) generate over €60 000 / year in subscription revenue alone. This is the 2027 target. Conservative.

Demand signal. The current public surface receives between 5 000 and 15 000 visits / month, with engagement concentrated on docs, pricing, and product tour pages. Community channels show consistent inbound from individual professionals asking for early access.

Distribution is product-led. No paid advertising budget at this stage. Customers reach the product through five channels.

1. Public content. Essays and product notes on X across two accounts (@Zer0H1ro founder voice and @odei_ai brand voice). Consistent reach in technical communities, with individual posts regularly exceeding 100 000 impressions.
2. Direct referrals. Existing users invite peers. Each paid subscriber receives credit for inviting another paying subscriber. Compounding distribution at near-zero cost.
3. Developer ecosystem. Public API and documentation at api.odei.ai/docs attract developers. A subset introduce ODEI.AI to their end users.
4. Conference presence. Selective appearances at AI / developer conferences within the EU. Practitioner audiences, not enterprise procurement. Modest travel budget.
5. Local Hungarian network. Budapest software / startup community participation. Informal meetups, founder community.

Customer acquisition cost. Bounded by the cost of producing content and operating the website. Included in operating costs above.

Operational requirements are intentionally modest, listed below in order of materiality.

Why Hungary. EU legal framework for online services to EU customers. Favourable cost structure (rent, accountancy, software, insurance) for an early-stage software business. Active Budapest software / developer community as a first source of users and feedback. Personal residence supports both business and founder's continued integration into the local ecosystem.

I am Anton Illarionov. I have 13+ years of experience building software startups, including two successful companies shipped to production.

Most recently I was Co-founder & Builder at Aegis (Feb 2024 – Jul 2025), where I architected YUSD, a Bitcoin-backed stablecoin protocol with real-time Proof-of-Reserves and intrinsic yield generation. The protocol scaled to $40M+ TVL in production, passed Hacken and Sherlock audits on the first cycle, and was deployed under a Cayman Foundation + BVI operating structure.

Earlier, I served as Entrepreneur in Residence at Google for Startups, where I worked with early-stage companies on product strategy and go-to-market execution.

Since July 2025 I have been the operator and sole owner of ODEI.AI — The Human API Company. The product, codebase, and customer relationships are mine. Day-to-day operations: product development, customer support, content, financial administration, strategic decisions.

Public profile: antonillarionov.com. The sole-proprietor registration was completed on February 25, 2026 and has been continuously active, with quarterly filings on schedule.