Medium potentialLow license · Apache-2.0Score 8/10
RLT Forecast API
Fast, looped transformer for time series
A hosted inference API that lets you run the Recurrent Looped Transformer on your time series data, delivering accurate predictions with low latency.
Tech stack
FastAPI + PyTorch + Docker + AWS SageMaker + Stripe + Clerk
BYOK model
User must provide their own AWS credentials for GPU instance
Freemium model
Free tier: 1,000 inference requests/month, 1 GPU instance; Pro tier: unlimited requests, priority GPU, model fine‑tuning, SLA.
Deployment complexity
Medium — Requires GPU for inference but can be hosted on managed services such as AWS SageMaker or GCP AI Platform.
SEO keywords
recurrent looped transformerRLT inference APItime series transformerreal-time fraud detection modellow latency transformer inference
Why it qualifies
- ✓High star count indicates strong community interest
- ✓Apache license permits commercial use without copyleft constraints
- ✓Niche architecture with no direct SaaS competitor
Build this