Self-host doods2
tf-v2.7.0-depsObject detection REST API for images and video streams
- Stars
- 278
- Port
- 8080 (HTTP REST API and web UI)
snowzach/doods2What is doods2?
DOODS2 (Dedicated Open Object Detection Service) is a Python rewrite of the original DOODS project that provides a REST API for object detection in images and video streams. It supports multiple detector backends including TensorFlow Lite, TensorFlow 1.x frozen graphs, PyTorch/YOLOv5, DeepStack, and Ultralytics YOLO models. The service can process single images via base64 or URL, and handle live video streams (RTSP, HTTP) with real-time detection feedback via MJPEG or WebSocket. It includes a simple web UI for testing, supports hardware acceleration via NVIDIA GPUs and Google Coral EdgeTPU, and offers an MQTT mode for asynchronous integration. Configuration is done via YAML file or environment variables. The project is distributed as multi-architecture Docker images (amd64, arm64, armv7) with GPU-enabled variants.
Should you self-host doods2?
doods2 scores 47/100 on selfhost.directory (how scoring works) and ranks #2 of 2 in Computer Vision. It has 278 GitHub stars.
There was no stable release in the past 12 months, so check the repository's recent activity before you rely on it.
Setup is moderate: expect some configuration beyond a single command. An official Docker image is available. It is MIT-licensed and written mostly in Python.
Its closest alternative on selfhost.directory is double-take (51/100, 701 stars). It scores higher overall, so compare the two before you pick. See doods2 vs double-take side by side.
Features
- Annotated Image Output
- Configurable Detection Thresholds
- Google Coral EdgeTPU Support
- MJPEG Streaming Output
- MQTT Mode
- Multiple Detector Backends
- NVIDIA GPU acceleration
- Object Detection REST API
- Preprocessing Filters
- Region-based Detection
- Video Stream Detection
- WebSocket API
- Web UI for Testing
doods2 README
More in Computer Vision
double-takeDouble Take is a self-hosted, open-source web app that unifies facial recognition across multiple detectors and NVRs like Frigate.
LLM VisionSelf-hosted Home Assistant integration using multimodal LLMs to analyze camera feeds, images, videos, and Frigate events with intelligent notifications and event timeline
locate-anythingSelf-hosted web UI for NVIDIA LocateAnything-3B: find objects, text, and UI elements in images via plain-language prompts on your GPU.
RembgOpen-source, self-hosted background remover: a Python library, CLI and HTTP server that cuts subjects out of images using switchable ONNX models.
VisionStateSelf-hosted Home Assistant addon that teaches local AI to recognize states, detect objects, and read meters from camera feeds — all on your CPU.
DeepSeek-OCR-WebUISelf-hosted, open-source OCR web app and REST API that converts images and PDFs to Markdown using the DeepSeek-OCR-2 model, with 7 modes and batch processing.
0 comments