Market data foundation
Normalize CSV imports, provider polling, live candles, and Go worker ingestion into one auditable Postgres model.
Final-candle analysisProvider healthGap recovery
API https://ai-trading-agent-api.vercel.app
A production-minded starter kit for market data, deterministic intelligence, grounded AI explanations, paper-only agent plans, and daily review workflows.
Readiness
92%
Review first
07
Data health
GOOD
Deterministic engines classify. AI explains supplied evidence. Humans review the result.
No broker executionWhat you get
Use the full opinionated stack or replace modules behind typed boundaries as your product evolves.
Normalize CSV imports, provider polling, live candles, and Go worker ingestion into one auditable Postgres model.
Build signals from typed indicators, patterns, quality gates, market regimes, evidence, and reproducible engine runs.
Layer explanations and paper-only planning over persisted evidence without letting a model rewrite source signals.
Start with Neon-backed accounts, workspace RBAC, job queues, worker health, audit trails, and fail-soft dashboards.
Architecture
Python remains the product brain. Go handles bounded ingestion work. Neon stores the auditable source of truth.
Quick start
Bring up the full development stack, apply the schema, and seed deterministic defaults.
$ make dev
$ make migrate
$ make seed