Self-Hosted Facial Recognition API
Budget / Salary£20–250
TypeFreelance project
LocationRemote
Posted2 hours ago
## Self-Hosted Facial Recognition API – Python / PHP
We need a developer to build a **self-hosted facial recognition system** using available open-source/pre-trained libraries.
We do **not** want to use paid services such as Azure Face API, AWS Rekognition, Face++ etc.
The system should run on our own server and provide a simple **web service/API** which our existing PHP application can call.
Main requirements:
* Enrol a person using one or more face images
* Preferably work even when only **one enrolment image** is available
* Recognise an enrolled person from a submitted image
* Return Person ID / match result / confidence score
* Ideally support multiple faces in one image
* Store facial embeddings/templates efficiently
* Fast response and suitable for many enrolled users
* Simple REST/API interface using POST requests
* Provide PHP example code for sending images and receiving JSON response
* Full source code and installation instructions
You may use suitable technologies such as Python, OpenCV, ONNX, InsightFace/ArcFace, FAISS or other appropriate open-source solutions.
We are **not asking to create an AI model completely from scratch**. Using and configuring suitable existing pretrained models/libraries is acceptable.
The final solution must operate independently on our server without any mandatory per-image or monthly facial-recognition subscription.
When bidding, please briefly mention:
1. Which facial recognition library/model you propose.
2. Whether you have done similar work.
3. Whether it can work with single-photo enrolment.
4. Approximate development time and cost.
We need a developer to build a **self-hosted facial recognition system** using available open-source/pre-trained libraries.
We do **not** want to use paid services such as Azure Face API, AWS Rekognition, Face++ etc.
The system should run on our own server and provide a simple **web service/API** which our existing PHP application can call.
Main requirements:
* Enrol a person using one or more face images
* Preferably work even when only **one enrolment image** is available
* Recognise an enrolled person from a submitted image
* Return Person ID / match result / confidence score
* Ideally support multiple faces in one image
* Store facial embeddings/templates efficiently
* Fast response and suitable for many enrolled users
* Simple REST/API interface using POST requests
* Provide PHP example code for sending images and receiving JSON response
* Full source code and installation instructions
You may use suitable technologies such as Python, OpenCV, ONNX, InsightFace/ArcFace, FAISS or other appropriate open-source solutions.
We are **not asking to create an AI model completely from scratch**. Using and configuring suitable existing pretrained models/libraries is acceptable.
The final solution must operate independently on our server without any mandatory per-image or monthly facial-recognition subscription.
When bidding, please briefly mention:
1. Which facial recognition library/model you propose.
2. Whether you have done similar work.
3. Whether it can work with single-photo enrolment.
4. Approximate development time and cost.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.