Analyse Radio Signals at Cloud Scale
Upload Breakthrough Listen observation files. Get ML-classified signal candidates back in seconds. Powered by a Rust engine that runs 45x faster than turboSETI.
Cloud Processing Pipeline
From raw telescope data to classified signal candidates, fully automated in the cloud.
Rust-Powered Processing
The same compiled Rust binary that achieves 45x speedup over turboSETI runs in the cloud on ECS Fargate.
Taylor Tree + Chirp Search
O(N log N) de-Doppler search with Numba JIT fallback, plus Doppler acceleration detection across 28 matched filter templates.
ML Classification
CNN + Transformer classifier with 9 signal classes. Out-of-distribution detection flags anomalies not seen in training.
Intelligent RFI Rejection
Spectral kurtosis filtering + 27-source RFI database. Automatic terrestrial interference labelling before classification.
FITS Catalog Export
Standard astronomical binary table output with astropy. Cross-reference against SIMBAD, NVSS, FIRST, and ATNF catalogs.
Signal Persistence
Track candidate signals across multiple observations. Persistent signals get elevated priority and cross-epoch correlation.
Pricing
Start free. Scale when you need to. All prices in USD.
- Supports .h5 and .fil files
- 10 files per month
- 50 MB max file size
- Standard processing
- Web upload only
- 7-day result retention
- Supports .h5 and .fil files
- 500 files per month
- 1 GB max file size
- Priority processing
- REST API access
- FITS catalog export
- Signal persistence tracking
- 90-day result retention
- Region selection (on request)
- Email support
- 5,000 files per month
- 5 GB max file size
- Priority processing
- REST API access
- FITS catalog export
- 1-year result retention
- Bulk upload
- Team access (10 seats)
- Webhook notifications
- Multi-region deployment
- Priority support (24hr)
- Custom file volume
- 10 GB max file size
- Everything in Institution
- GPU processing
- Custom ML models
- Self-hosted option
- Multi-region deployment
- Unlimited seats
- Dedicated support (4hr)
Two Ways to Use
Upload via the web dashboard or integrate programmatically with the REST API.
Web Upload
Drag and drop your .h5 or .fil files directly in the browser. Watch results appear in real-time. No coding required.
REST API
Automate with 3 API calls. Get presigned upload URLs, submit jobs, and poll for results. Python SDK coming soon.
What the Cloud Provides
Every core MitraSETI Cloud feature, plus cloud-specific capabilities.
| Feature | Cloud | Local (Open Source) |
|---|---|---|
| Rust de-Doppler (45x speedup) | ✓ | ✓ |
| GPU Taylor tree (CuPy/Numba) | — | ✓ (if GPU available) |
| Chirp rate search | — | ✓ |
| Matched filter bank (28 templates) | — | ✓ |
| ML classification | ✓ | ✓ |
| Web upload (drag & drop) | ✓ | — |
| Auto-scaling (100+ files parallel) | ✓ | — |
| Email/SMS alerts | ✓ | — |
| REST API | ✓ | ✓ (local) |
| Results dashboard | ✓ | ✓ (local) |
| Desktop app (PyQt5) | — | ✓ |
| Zero infrastructure setup | ✓ | Requires Rust toolchain |
Engineered by Deep Field Labs
The same engineering that powers this platform is what we deliver to institutions and enterprises worldwide.
MitraSETI Cloud is designed and maintained by Deep Field Labs. We build cloud platforms and ML infrastructure that scale from prototype to production.
Get in Touch
For institutions and enterprises: We deliver the same architecture patterns shown here, including serverless orchestration, Rust + ML pipelines, and Terraform automation, as consulting engagements.