Skip to main content
Back to selected work

CryptoAI Trader

AI-assisted trading workflows in one SaaS product.

A multi-service application connecting exchange data, deterministic strategy automation, AI-assisted analysis, subscriptions and scheduled operations.

View live product
CRYPTOAI / SIGNAL ENGINEMARKET LIVE
01Market data02AI signal03Risk gate04Execution
BTC/USD+2.84%SIGNAL VALIDATED
CONFIDENCE 87%RISK CONTROLLED
Ownership
Product architecture · data and AI workflows · SaaS engineering
Platform
Next.js · Supabase · Stripe · Coinbase · Kraken
Domain
Crypto automation SaaS
Status
Production
01

Context and operational problem

Market data, strategy research, exchange connectivity, portfolio history and subscription operations usually live in separate tools. That fragmentation makes both the trading workflow and the commercial product harder to operate.

CryptoAI Trader separates market ingestion, strategy logic, AI-assisted context, risk controls and execution into focused service boundaries, then brings them together in one authenticated SaaS experience.

02

System architecture

MARKETCoinbaseKrakenMarket data services
INTELLIGENCEGrid / DCA strategiesAI-assisted contextDeterministic risk rules
EXECUTION & PRODUCTScheduled executionPortfolio & historyStripe SaaS
03

What was built

  • Grid and DCA strategy workflows
  • Coinbase and Kraken service integrations
  • AI-assisted market analysis
  • Portfolio and trade history
  • Stripe subscription lifecycle
  • Authenticated scheduled market operations
04

Engineering decisions

01

Separate analysis from execution

AI-assisted context is kept distinct from deterministic strategy and risk rules so generated output is not the execution authority.

02

Protect exchange credentials

Sensitive settings remain server-side, encrypted and isolated by user rather than entering the browser runtime.

03

Orchestrate scheduled work

Authenticated cron routes coordinate analysis, strategy evaluation and portfolio snapshots across focused services.

05

Evidence and validation

Coinbase and Kraken servicesTwo exchange boundaries implemented independentlyVitest strategy testsDeterministic trading rules validated outside the UIStripe billing lifecycleSubscription state connected to product access
06

What I learned

The strongest AI products define where model assistance stops. In this system, AI adds context while deterministic services retain responsibility for strategy constraints, risk checks and execution. That boundary is as important as the model itself.

Have a complex workflow to simplify?

Start a conversation