# jakowenko/double-take

Unified UI and API for processing and training images for facial recognition.

Repository: https://github.com/jakowenko/double-take
Canonical: https://ross.abutalabs.com/products/double-take
Homepage: https://hub.docker.com/r/jakowenko/double-take
Language: JavaScript
License: MIT
License Family: permissive
Topics: frigate, mqtt, face-recognition, compreface, facebox, deepstack, home-assistant, home-automation, room-presence, rekognition
Last push: 2025-08-18T20:14:39+00:00

## Health v2 (maintenance only)
Score: 40/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 37, release rhythm 11, longevity 100
- inputs: {"age_days": 2001, "days_push": 380, "days_rel": 380, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1436, forks 146 (observed 2026-08-28T04:04:43.684303+00:00)

## What it is
Double Take is a self-hosted Docker application providing a unified UI and API for facial recognition. It abstracts multiple face detection services (CompreFace, DeepStack, Facebox, Amazon Rekognition) behind one interface, processing images from NVRs like Frigate and publishing results over MQTT.

## Use cases
- run facial recognition on security camera snapshots
- unify multiple face detection backends behind one API
- identify people from Frigate camera events
- train face recognition subjects from a web UI
- publish face match results to MQTT for Home Assistant
- detect room presence based on who is seen on camera

## When to choose
- you use Frigate or another NVR and want person identification
- you want to swap or combine face recognition engines without changing your automation
- you want a ready-made UI for training and managing face subjects
- you want MQTT/REST integration with Home Assistant

## When to avoid
- you need general object detection rather than face recognition
- you want a standalone face recognition engine rather than an orchestrator
- you cannot run Docker containers
- you need real-time video stream analysis rather than snapshot processing

## Facets
- artifact type: application
- maturity: active
- function: computer-vision, image-processing, api-framework, web-framework, machine-learning, middleware
- domain: computer-vision, image-processing, self-hosted, iot
- platform: self-hosted
- tags: face-recognition, frigate, mqtt, home-assistant, nvr, compreface, deepstack, facebox, rekognition, room-presence, home-automation, docker, web-server, linux

## Member repositories
- jakowenko/double-take (main) score 40

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:43.684303+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:36:48.037569+00:00, confidence not recorded.
  - readme: https://github.com/jakowenko/double-take (fetched 2026-08-28T04:04:43.684303+00:00, sha c24e9f912d94)
  - homepage: https://hub.docker.com/r/jakowenko/double-take (fetched 2026-08-29T11:47:53.254592+00:00, sha e3ef3cb1ac37)
  - site_page: https://docs.docker.com/docker-hub (fetched 2026-08-29T11:47:53.264162+00:00, sha caaa18892709)
  - site_page: https://www.docker.com/docs (fetched 2026-08-29T11:47:53.266085+00:00, sha aeada2272712)
  - site_page: https://www.docker.com/company (fetched 2026-08-29T11:47:53.267899+00:00, sha 05ab59b2bc29)
- Data as of 2026-08-30T08:39:29.467469+00:00.