case · 002
Solana trading platform, end-to-end.
A microservices platform for trading Solana tokens end-to-end: token-discovery service pulls signals from Telegram channels and DexScreener, copy-trading service mirrors selected wallets, and a trade-execution service routes orders through Jupiter. Each module ships as its own Express + Prisma service with its own port and its own schema, coordinated behind a Next.js operator console.
fig. 0.1
Cover image · microservice topology
Solana memecoin windows move in minutes — by the time a single monolith finishes ingesting a Telegram signal and running a rug-check, the trade window is gone. Splitting into four focused services (alpha-radar signals, copy-pulse mirroring, trade execution, shared types) let signal ingestion run at message rate while execution stayed independent.
Every signal and every order is logged for post-trade review and audit. Rug-check failures are surfaced to the operator console within seconds of the source contract going live, and the trade-execution service refuses to route orders against contracts that fail risk thresholds without an explicit override.
microservices in production
signal sources wired (Telegram, DexScreener, Jupiter, Rugcheck)
audit memos recorded for every order
The shared types workspace holds the contract, signal, and order schemas. Both data and event semantics flow through it, so a schema change in alpha-radar is caught by the CI build of every other service before it can drift.
Backpressure is per-service: the alpha-radar Redis queue bounds memory; the copy-pulse execution service honours Telegram rate limits and batched rebalances; the trade execution service degrades to dry-run when Jupiter aggregator quotes stale or slippage exceeds threshold, never placing an unprofitable order silently. An operator dashboard renders all of this live via WebSocket.
/ stack
- `@solana/web3.js` (Solana RPC)
- Jupiter aggregator
- Rugcheck risk API
- Next.js 14
- Express
- Node.js
- TypeScript
- PostgreSQL (Prisma)
- Redis
- Telegram Bot API
- Google OAuth
next case study
Smart-contract risk monitor, 16 chains..
A continuously-running feed poller that watches 16 chains' Blockscout v2 explorers for newly verified smart contracts, runs each one through Slither (static analysis) plus a multi-label ML risk classifier, and persists contracts and findings to PostgreSQL with a per-chain cursor that's decoupled from analysis so crashes never cause re-fetching or data loss. Served to operators through an OAuth-gated dashboard with contract listings, findings, chain coverage, and an analysis console.
/ start a project