Skip to main content
FluxState Edge Hero Banner

🦅 FluxState Edge SDK

Privacy-Preserving Contextual Edge Video Analytics for Enterprise Security

License: MIT Tests: Passing


FluxState Edge is an extensible, camera-agnostic video analytics SDK designed for direct integration into proprietary security backends. It processes RTSP streams locally on the edge, extracting behavioral metadata via object detection, 3D skeletal posing, audio processing, and VLM semantic reasoning with our v1.1 Adaptive Intelligence layer.



🚀 Quick Install

pip install fluxstate-security

🎥 Installation Demo

See the SDK installation and verification in under a minute.

▶️ Watch Installation Demo (Google Drive)


⚡ System Capabilities

Capability Technical Implementation
🧠 Agentic VLM Reasoner Runs Vision-Language Models (e.g., Qwen2.5-VL) locally via MLX on Apple Silicon unified memory for deep contextual scene understanding. Bypasses cloud APIs entirely.
🔄 Adaptive Intelligence (v1.1) Seamlessly loads and unloads MLX models (MLXModelPool) to manage unified memory. Provides a Feedback API to suppress false positives on-device without retraining.
📚 Semantic Retrieval (RAG) Embeds scene telemetry and queries an on-device ContextLedger using cosine similarity to inject historical context into the VLM prompt.
⏱️ Episodic Memory Buffers the last 10 minutes of scene states (EpisodicMemoryBuffer) with O(1) operations, allowing delayed asynchronous operator feedback.
🎯 Tactical Visual Grounding Automatically injects high-contrast red bounding boxes over anomalies. This guides the VLM toward the detected region, reducing irrelevant reasoning and improving contextual grounding.
🛡️ Privacy-by-Design Actively zeroes out image buffers post-inference via C-level memset. Designed with privacy-first principles to prevent sensitive pixel data from lingering in the system heap.
🗄️ Temporal Forensics Behavioral anomalies are serialized into a local SQLite database (fluxstate_security/core/forensics.py), creating a searchable text-based ledger of physical events.
📹 Hardware Agnostic Ingests existing IP cameras via standard RTSP URLs. No proprietary recording hardware required.
🐳 Edge Containerization Ships with a highly optimized Dockerfile for enterprise edge deployments (Kubernetes/Docker Swarm), permanently locking native OS dependencies.

🖼️ Architectural Vision & Use Cases

FluxState Architecture Diagram

System Overview

Sub-Pipelines

Adaptive Pipeline RAG Flow Feedback Loop

Tech Stack

🛠️ Minimal SDK Integration

FluxState is designed to fade into the background. Drop it into your existing backend and attach a webhook.

import time
from fluxstate_security import FluxStateNode

# 1. Initialize the SDK
sdk = FluxStateNode()

# 2. Define your integration hook
def handle_threat(event_payload):
    print(f"\n[INTEGRATION BUS] Escalating to VMS...")
    print(f"Target Identity: {event_payload['entities']}")
    print(f"Behavioral Vector: {event_payload['context_log']}")
    print(f"Feedback Event ID: {event_payload.get('event_id', 'None')}")

# 3. Bind the hook
sdk.on_threat_detected = handle_threat

# 4. Deploy Headlessly (Runs as a background daemon)
sdk.start_headless_daemon()
sdk.start_api_server(port=8000) # Starts the JSON API & Feedback loop

try:
    while True: time.sleep(1)
except KeyboardInterrupt:
    sdk.stop() # 5. Clean shutdown

📦 Deployment Strategies

Option A: Enterprise Docker Deployment (Recommended)
For production environments, use the provided Docker container to guarantee system dependencies (Tesseract, PortAudio) are perfectly locked across the cluster.
docker build -t fluxstate-security .
docker run -d --name fluxstate-security fluxstate-security
Option B: Local Python Development
pip install fluxstate-security

# macOS dependencies
brew install tesseract portaudio

# Linux dependencies
sudo apt-get install tesseract-ocr libportaudio2 libportaudiocpp0 portaudio19-dev

🧪 Forensics & Testing

FluxState ships with an automated pytest suite covering the Forensic SQLite ledger, semantic retrievers, and JSON intelligence policies.

pytest tests/

📡 API Reference

POST /api/v2/feedback

Submit operator feedback to adapt the system. Payload:

{
  "event_id": "uuid-string",
  "human_label": "FALSE_POSITIVE",
  "operator_id": "guard-123",
  "operator_role": "admin"
}

For a deeper dive into the threading model, VLM orchestration, and the SQLite schema, see architecture_adaptive_edge.md.

Metadata

Release files for fluxstate-security 1.1.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for fluxstate-security 1.1.0
File Size Uploaded
fluxstate_security-1.1.0.tar.gz 41.9 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for fluxstate-security 1.1.0
File Interpreter ABI Platform
fluxstate_security-1.1.0-py3-none-any.whl Python 3 none any Details

Total release size: 77.1 kB

Release files / fluxstate_security-1.1.0.tar.gz

Download URL fluxstate_security-1.1.0.tar.gz
Size 41.9 kB
Tags Source
SHA-256 checksum
How to use checksums
c16accb06da9d3907455442cb0c9bdb23b204430f87781996c808ce61954657b
BLAKE2b-256 checksum
How to use checksums
1882ced3dad12c4fda34812c4b7cf69f79137d10531446fc90f5f64ab6b68098
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 Jul 6, 2026.

Transparency log

Release files / fluxstate_security-1.1.0-py3-none-any.whl

Download URL fluxstate_security-1.1.0-py3-none-any.whl
Size 35.1 kB
Tags Python 3
SHA-256 checksum
How to use checksums
f16e8f1bff4fcffe0b8617133b1a24aafeb62b5681f7595573bf6f42c9c8381e
BLAKE2b-256 checksum
How to use checksums
8f3172c9c3aa56fedd0c44cef75a8d7238ca7fd02b64d548bc53689717abb50b
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 Jul 6, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

1.1.0 This release

2 release files

1.0.2

2 release files

1.0.1

2 release files

1.0.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page