---
title: "Custom APIs"
description: "How to expose object metadata to non-Python 3rd party applications by mounting a custom Flask app alongside corridor-api."
---

This helps create custom APIs to use for 3rd party integrations by using the [`corridor` package](/technology/customization/fetching-metadata/corridor-package/).
The user would define additional flask-applications which would be mounted along with Corridor API.
This option is useful if the 3rd party application is not python based, and therefore can't use the `corridor`
package directly.

## Utility

There are scenarios when the user wants to use a 3rd party application which fetches the details
about the objects on the platform. To accomplish that, the user can create a simple `Flask` application,
and register the corresponding APIs. These APIs would use the `corridor` package to fetch the
required information. This new `FLASK` app would be mounted along with the `corridor-api` application,
and we would internally register those API endpoints along with the rest of `corridor-api` endpoints.

## Example

The example focuses on creating a `GET` api to fetch information about all the models to which a given user has
access to. The 3rd party application can specify the `username` as part of the headers with the key: `x-user-name`.
The endpoint in this case would be: `http://localhost:5000/{CUSTOM_APP_NAME}/models`, assuming `corridor-api` is running
on port: 5000.
The user can optionally provide `prefix` kwarg in the configurations, in which case the endpoint would become:
`http://localhost:5000/{prefix}/models`

```python
CUSTOM_APP_NAME = 'customtool'

from flask import Flask, request

custom_app = Flask(CUSTOM_APP_NAME)


@app.route('/models', methods=['GET'])
def models():
    import corridor
    from corridor_api import AppUser

    username = request.headers.get('x-user-name')
    with AppUser(username):
        models = corridor.Model.models
        models_info = {}
        for model in models:
            model_obj = corridor.Model(name=model)
            model_id = model_obj.id
            models_info[model_id] = {
                'modelId': model_id,
                'modelName': model_obj.name,
                'modelVersion': model_obj.version,
                'modelType': model_obj.type,
                'outputAlias': model_obj.output_alias,
                'createdBy': model_obj.created_by,
                'current_status': model_obj.current_status,
            }

        return {'result': models_info}

```

:::note
The user, for which we want to fetch the information, needs to be passed in the headers.
`AppUser` sets the user context for us, so that only the information accessible to that user is fetched.
:::

## Configurations

Custom APIs related configurations need to be set in [`api_config.py`](/technology/installation/configuration-setup/api/) along with other configurations.

```python
CUSTOM_APIS = {
    CUSTOM_APP_NAME: {'prefix': 'optional_prefix_to_add_to_the_endpoint', 'app': custom_app},
}
```