> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dojah.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Face Authentication

> Check who someone is by looking at their face, and keep a record of every check your business makes.

Face Authentication is a set of tools that let a business check who someone is by looking at their face and keep a record of every check it makes. A business can add a person's photo to its own records, check a new photo against a photo already on file, or check two brand-new photos against each other. Every one of these checks is logged, so the business can look back at what happened.

## The problem it solves

Businesses that verify identity online (opening an account, approving a loan, letting someone back into a service) need a fast, reliable way to confirm a person is who they say they are, using nothing but a photo. Doing this by eye does not scale and is easy to get wrong. Face Authentication does the comparison automatically and gives a confidence score, a clear match or no-match result, and a record of the check for later reference.

## What you can do with it

There are three things a business can do with Face Authentication. Each one is logged separately.

| Action | What it does | When a business would use it |
| - | - | - |
| Register (also called Enrol) | Adds a person's photo to the business's own records, under a named group and a reference ID | Onboarding a new customer or employee whose face the business wants on file for future checks |
| Authenticate (also called Verify) | Compares a new photo against a photo already on file for a specific person and returns a match or no-match result with a confidence score | Confirming a returning user is the same person who was originally registered, such as at login or before a sensitive action |
| Compare (also called Match) | Checks two brand-new photos against each other directly. Neither photo needs to already be on file | A one-off check between two images, for example confirming a submitted photo matches an ID photo, with nothing saved to a person's record beforehand |

Every check a business runs, whichever of the three it is, is always tagged with a reference ID. This is a unique label the business gives (or the system generates for) that check, so it can find that exact record again later and investigate if something goes wrong.

## Where to find it in your dashboard

Face Authentication has its own place in the Dojah dashboard, under the "Verify" section in the left-hand menu, alongside Individual and Business. Every Dojah customer can see it. There is nothing to request or turn on.

The dashboard does not run a face check itself. There is no screen for uploading a photo and getting a Register, Authenticate, or Compare result back. That work happens entirely through the [API](https://docs.dojah.io/api-reference/face-auth/collections/create-collection.md), called from the business's own integration. What the dashboard gives you is everything around that work: organizing it beforehand and reviewing it afterward.

### The Collections page

* Create a collection. The only thing needed is a name.
* Search and filter existing collections by date.
* See the total number of collections on the account.
* Open a collection to see when it was created, its ID, and how many faces are registered in it.
* Delete a collection.

![The Face Authentication Collections page listing named collections](https://mintcdn.com/dojah/6vNPjl5Pl_mFf767/images/dashboard-guide/face-auth-collections.png?fit=max\&auto=format\&n=6vNPjl5Pl_mFf767\&q=85\&s=535f3ceab4ed1d995b3058ae7d6e00e3)

### The Face Logs page

* A read-only history of every check the account has made: Register, Authenticate, and Compare.
* Each entry shows its Collection (where one applies), Reference ID, Action, Date Created, Status, and Confidence score.
* Filter by date, Collection, action, or status.
* Summary numbers at the top: how many checks in total, how many succeeded, and how many failed.
* Search for a specific check by its Reference ID.
* Export the log as a file.

![The Face Logs page with summary counts and a filterable history of checks](https://mintcdn.com/dojah/6vNPjl5Pl_mFf767/images/dashboard-guide/face-auth-logs.png?fit=max\&auto=format\&n=6vNPjl5Pl_mFf767\&q=85\&s=60a215edaf350d20e7cfe1d51e28f9d6)

> [!NOTE]
>
> The API is where the actual verification work happens, submitting a photo and getting a match or no-match result back. The dashboard is where a business organizes that work into collections beforehand and reviews what happened afterward.

## Collections and Reference IDs

A Collection is a named group of registered faces; for example, a business might have one Collection for customers and a separate one for staff. A person's photo lives inside a Collection once it has been registered.

A reference ID is the unique label attached to one specific person's record inside a Collection, for example, a customer ID or account number the business already uses. Authenticate looks a person up by their Reference ID inside a Collection to find the photo it should compare against.

Compare does not use a Collection at all, because neither of its two photos needs to be registered anywhere first. It still carries a Reference ID, for the uniqueness and investigation reasons noted above, but there is no group of faces behind it.

## How it works

1. **Register a person.** The business submits a photo, along with a Collection name and a Reference ID. That photo is now on file.
2. **Time passes.** The person comes back, for example, to log in, complete a transaction, or pass a checkpoint.
3. **The business submits a new photo.** Along with the same Reference ID and Collection used at registration.
4. **Face Authentication compares the two.** The new photo against the one on file, and returns a match or no-match result, plus a confidence score.
5. **The business acts on the result.** Let the person through, flag the account, ask for another check, whatever fits their process. The check is now in their history log.

## Where businesses use it

A few examples of how this shows up in practice. These are illustrations of the kind of use case Face Authentication fits, not real customer stories.

* A lending app registers a customer's photo at signup, then re-checks their face against it before releasing a large withdrawal.
* A ride-hailing platform registers each driver's photo once, then asks for a fresh photo at the start of a shift to confirm the right driver is behind the wheel.
* A background-check service compares a submitted selfie against a photo on a government ID, without ever storing either photo permanently.

## How it helps

* **Cuts manual review.** A business no longer needs a person to eyeball two photos side by side.
* **Gives a clear, repeatable result.** A confidence score and a match or no-match answer, not a judgment call.
* **Keeps a record.** Every check is logged with a timestamp and a reference ID, so a business can answer "what happened here" later.
* **Fits different situations.** Register-then-Authenticate for people the business already knows, Compare for a one-off check with nothing on file.

## Reading a result

| Term | What it means |
| - | - |
| Match | The two photos compared (or the new photo and the one on file) are judged to be the same person |
| No match | The two photos are judged to be different people. This is a normal, successful result, not an error |
| Confidence score | A number showing how sure the system is about its match or no-match decision. It applies to Authenticate and Compare, not to Register, since Register has nothing to compare against yet |
| Check failed | The check itself could not run to completion, for example, because no photo was submitted or the system could not find the record it needed. This is different from a check that ran and simply found no match |


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.