Self-hosted crypto payment gateway for consultants and digital-service operators. Accept BTC, ETH, and USDT direct to your wallets — no KYC, no processor cut, no SaaS middleman.
Stripe charges 2.9% per transaction and demands KYC from sellers. Crypto-native buyers walk past anything that asks for either, and a meaningful slice of every consultant's TAM walks with them. CryptoPay lets a service operator accept BTC, ETH, and USDT direct into their own wallets, with no processor cut, no platform account, and no funds held in third-party custody.
A consultant doing $50,000 a year in service revenue hands roughly $1,500-$2,000 to card processors every year, then waits 2-7 business days for each payout to settle. Stripe takes 2.9% plus $0.30 per transaction, PayPal layers on a 0.5% currency-conversion margin on cross-border sales, and both reserve the right to freeze a seller's account on suspicious volume patterns, a routine occurrence for anything adjacent to Web3, OSINT, or privacy-sensitive work.
Worse, both processors silently exclude the buyers most likely to pay premium rates. DeFi participants, Web3 builders, and serious crypto holders refuse to liquidate ETH to fiat for a $499 consulting session. They want the option to pay in crypto, and they bounce from checkouts that don't offer it. The standard alternative, SaaS crypto-payment processors like Coinbase Commerce, BitPay, and NOWPayments, reintroduces 1-1.5% platform fees, ties the seller to a custodial wallet that can be frozen, and re-adds a third-party logo to the checkout that the buyer was specifically trying to avoid.
A buyer lands on the store, picks a service, fills in their name plus an email or Telegram handle, and chooses BTC, ETH, or USDT. They see a unique wallet address with a QR code and the exact amount to send. Once the payment hits the blockchain and confirms, they see the operator's custom fulfillment message — a calendar link, an access code, a download URL, whatever the operator wants to deliver. End-to-end, the buyer never creates an account, never gives a card number to anything, and never sees a third-party logo on the checkout page.
Each order generates its own wallet address, so there is no amount-matching problem — the system knows exactly which payment belongs to which buyer. A background monitor polls BlockCypher, Etherscan, and TronGrid every 60 seconds, watches for incoming transactions, counts confirmations, and flips the order to confirmed once it clears the configured threshold (BTC 2 confirmations, ETH 12, USDT 19). The buyer's payment page updates live over a WebSocket — they do not have to refresh, and they do not get stuck wondering whether the payment landed. An admin panel gives the operator full visibility: revenue by currency, revenue by service, the full order pipeline, wallet balances, and one-click sweep to a corporate wallet.
The whole stack drops onto a low-end VPS as three Docker containers. No SaaS account, no platform terms of service, no processor that can suspend the operator's livelihood.
The backend is FastAPI on Python 3.11 with SQLAlchemy 2.0 async over PostgreSQL 15 — proven foundations that will still be running in five years. Blockchain integration is split across three RPC providers (BlockCypher for BTC, Etherscan plus Alchemy RPC for ETH, TronGrid for USDT TRC-20), each isolated behind its own monitor class so swapping a provider is a one-file change. Wallet private keys are encrypted at rest with Fernet before they ever touch the database. The frontend is Next.js 15 and React 19 with TailwindCSS, deployed as a standalone build. The whole stack stands up on a low-end VPS as three Docker containers, with interactive setup, start, monitor, and cleanup scripts so an operator without DevOps background can run it.
CryptoPay is feature-complete across the five build phases in the project's implementation plan: backend scaffolding, blockchain monitors, admin panel, public storefront, and DevOps tooling are all working. Roughly 6,800 lines of code split between FastAPI Python (around 3,000 LOC, 12 service files plus three blockchain monitors) and Next.js TypeScript (around 3,000 LOC, including a 1,579-line admin panel covering dashboard, orders, services, wallets, and analytics tabs). The codebase was assembled as a tight, single-purpose tool over February-March 2026, last touched 2026-03-08, and is positioned as a deployable product for a single operator rather than a multi-tenant SaaS. Honest framing: this is BETA-stage operator tooling, not a publicly hosted reference instance — the deliverable is a working tool that a buyer drops into their own infrastructure, not a service the architect operates on their behalf.
The next milestone is packaging CryptoPay as a turnkey deployment — wrapping the existing setup wizard into a 10-minute provisioning experience on common VPS providers, hardening the admin panel for non-technical operators, and adding the optional payment methods the market most often asks for (additional ERC-20 stablecoins, Solana USDC, and Lightning Network for sub-$50 services where on-chain fees would dominate). The medium-term direction is a managed-deployment offering for sellers who want the no-middleman economics without standing up a VPS themselves — same self-custody guarantees, none of the operational burden.
CryptoPay is available in three engagement modes. A consultant or service operator can commission a custom deployment — branded, configured to their store, with corporate wallets and fulfillment flows tuned to their offerings. A team already running a service-sales site can extend their existing stack with this payment layer alongside their current checkout. And operators standing up their own crypto-payment rails can engage in strategic advisory on architecture, blockchain-monitor design, and the trade-offs between self-hosted and SaaS crypto-payment processing. Reach out via sintegrium.io or LinkedIn for a 30-minute scoping call.
Built by Yurii Staryk · Solution Ecosystem Architect
Self-hosted YouTube-to-text pipeline — faster-whisper runs on a home GPU inside Docker, callable from any laptop on the LAN. Own your transcripts, no API fees, no rate limits, 99+ languages.
Self-hosted LAN clipboard for engineers running multiple machines on one network — paste text or files on one device, grab on another, with auto-expiring slots for API keys and zero internet dependency.
A desktop wizard that authors complete cognition bundles — identity, skills, credentials, runtime memory — for multi-agent AI teams, sealed with AES-256 and ready to deploy to any runtime.
