rf-direction-finding
Hardware-agnostic RF direction finding, signal classification, and phased-array research.
What this is
rfdf is a Python platform for RF direction-finding (DOA), signal classification, and
phased-array experimentation. It is hardware-agnostic by design: every SDR, rotator,
antenna geometry, and compute provider sits behind an abstract interface (the HAL). The
B210 + AntRunner reference set is the first concrete validation of the abstractions, not
the system itself. A user with a HackRF, an RTL-SDR, or just a directory of SigMF files
is a first-class user.
rfdf runs standalone without ai-orchestrator. Installing the [orchestrator]
extra makes it a consumer of ernesto01louis/ai-orchestrator,
adding evidence-bundle production, Hindsight memory writes, L5 vault notes,
planner-dispatched GNU Radio flowgraphs, and ntfy alerts.
Validate RF research workflows on canned data before buying €8500 of gear.
Status
v0.1.0 — first stable release (Stage 7: orchestrator integration + REST
API). Stage 7 closes the build: an optional, lazy-imported integration with
ai-orchestrator —
consumer registration, citation-grade evidence bundles, Hindsight memory
writes, L5 vault notes, planner-dispatched GNU Radio flowgraphs, and ntfy
alerts — plus a standalone REST API (rfdf api serve) that doubles as the
orchestrator's capability callback target. The platform remains
standalone-first: pip install rfdf is a complete DOA + classification +
capture platform with no orchestrator awareness; the [orchestrator] extra is
a bonus, never a baseline. Prior stages: v0.1.0-alpha reference hardware
backends (USRP B210, AntRunner, GRBL rails, RTL-SDR/KrakenSDR contrib),
v0.1.0-beta the operator-facing software ecosystem (Ansible / Kasm /
Guacamole / monitoring, hosted alongside the platform, never a dependency).
With v0.1.0 the public API is stable under SemVer. See
docs/standalone-vs-orchestrator.md,
docs/orchestrator/, and ROADMAP.md.
Install (zero hardware path)
pip install rfdf # standalone — complete without an orchestrator
pip install rfdf[orchestrator] # + optional ai-orchestrator integration
The base install pulls only platform-essential dependencies (NumPy, SciPy, Pydantic, Typer, Rich, sigmf, structlog, platformdirs). Zero RF or ML dependencies are required to use the platform. Hardware and compute providers live behind extras and entry points.
Install (with hardware / compute extras)
| Extras | What it pulls in |
|---|---|
[sdr-uhd] |
Ettus UHD (B210 / B200mini / X-series) |
[sdr-pluto] |
pyadi-iio (PlutoSDR) |
[sdr-soapy] |
SoapySDR (HackRF, LimeSDR, others) |
[rotator-hamlib] |
hamlib-py (Yaesu / SPID / generic rotctld) |
[ml] |
PyTorch + TorchSig + scikit-learn |
[ml-onnx] |
ONNX Runtime for portable inference |
[compute-runpod] |
RunPod GPU rental |
[compute-modal] |
Modal serverless GPU |
[compute-vastai] |
Vast.ai marketplace GPU |
[compute-skypilot] |
SkyPilot multi-cloud GPU |
[orchestrator] |
ai-orchestrator-client integration |
[api] |
FastAPI + uvicorn (REST + WebSocket) |
[antenna] |
scikit-rf + PyNEC for antenna modelling |
[dev] |
Test + lint + type-check toolchain |
[all-hardware] |
All SDR + rotator extras together |
[all-compute] |
All compute backends together |
[all] |
The kitchen sink (ML + API + antenna + hardware + compute + orchestrator) |
Documentation
| Doc | What's in it |
|---|---|
| VISION.md | Why this exists, the principles, when-in-doubt questions |
| ARCHITECTURE.md | Layered design, HAL contracts, backend discovery, config precedence |
| ROADMAP.md | Stages 1–7 with status + acceptance criteria |
| CLAUDE.md | Working conventions (used by AI coding assistants) |
| CONTRIBUTING.md | Dev setup, branch flow, commit conventions |
| SECURITY.md | Threat model, secret handling, EIRP enforcement |
| CHANGELOG.md | Per-release changes |
License + acknowledgments
Apache-2.0. Matches ai-orchestrator and aero-research-platform.
This project builds on the work of many upstream RF and ML communities. Specific attributions live in docs/acknowledgments.md once that document is populated (Stage 3+).
1 maintainer
Metadata
1 maintainer
Release files for rfdf 0.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| rfdf-0.1.0.tar.gz | 296.2 kB | With provenance | Details for rfdf-0.1.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| rfdf-0.1.0-py3-none-any.whl | Python 3 | none | any | Details for rfdf-0.1.0-py3-none-any.whl |
Total release size: 548.4 kB
Release files / rfdf-0.1.0.tar.gz
| Download URL | rfdf-0.1.0.tar.gz |
|---|---|
| Size | 296.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
bbc3622fa2793f7648462d04c09358eb4a2e17e0e30e4691c48331f34ef2a983
|
|
BLAKE2b-256 checksum How to use checksums |
794cf0b140766243ecf0e69c01d609847c6631cab1bf2f53ea5deb3187b93197
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on May 19, 2026.
Transparency logRelease files / rfdf-0.1.0-py3-none-any.whl
| Download URL | rfdf-0.1.0-py3-none-any.whl |
|---|---|
| Size | 252.1 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
054afb2c5aff30f18f1b7c4f8187802d607c80781fae04150e0b7305aa4a8c72
|
|
BLAKE2b-256 checksum How to use checksums |
c7fada3c8f850d082802d9e09337bec8f5229de40fe652e5bb82a605fd7d5fb5
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on May 19, 2026.
Transparency logSecurity for rfdf 0.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
All 2 files in this release have been uploaded with provenance metadata.
Attestation for rfdf-0.1.0.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on May 19, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
