Integrations are configured in the nango.yaml file, which lives in your code base (learn more).

Example

nango.yaml
integrations:
  asana: # Integration ID.
    syncs:
      asana-fetch-tasks: # Sync name.
        description: | # Sync description.
          Some long description, some long description, some long description,
          some long description, some long description, some long description.
        output: AsanaTask # Synced model
        endpoint: # Generated endpoint to access the synced model.
          method: GET
          path: /tasks
          group: Tasks
        sync_type: incremental # Data is replaced ('full') or upserted ('incremental') on each sync execution.
        runs: every 30min # Sync frequency.
        input:  AsanaProject # Necessary input to execute the sync.
        scopes: # Required scopes.
          - scope1
          - scope2
        auto_start: true # If true, data syncing starts when a new connection is created.
        track_deletes: true # If true, automatically detects deleted records (for full refresh syncs only).
        webhook-subscriptions: # The types of webhook this sync handles.
          - event_type_1
          - event_type_2

    actions:
      asana-create-task:
        description: | # Action description.
          Some long description, some long description, some long description,
          some long description, some long description, some long description.
        output: AsanaTask # Returned model
        endpoint: # Generated endpoint to trigger to action.
          method: POST
          path: /tasks
          group: Tasks
        input: # Necessary input to execute the action.
        scopes: # Required scopes.
          - scope1
          - scope2

    on-events:
        post-connection-creation:
            - setup
            - moreSetup
        pre-connection-deletion:
            - cleanup

models:
  AsanaTask: # Schema for sync output AND action input/output.
    id: string # Required unique field for sync output models.
    project_id: string
    name: string
    completed: boolean
    created_at: date
    modified_at: date

  AsanaProject: # Schema for sync input.
    id: string

Integrations fields

Integration configuration fields are under integrations.<INTEGRATION-ID>.

syncs
object

Lists the syncs for a given integration.

actions
object

Lists the actions for a given integration.

on-events
object

Lists the events and scripts executed when those events occur.

Sync fields

Sync configuration fields are under integrations.<INTEGRATION-ID>.syncs.<SYNC-NAME>.

description
string
required

Describes the sync.

output
string
required

Defines the schema of the data you want to sync. References a schema from the Models section of this configuration file.

Each sync has a corresponding generated endpoint to fetch the data, configured in the endpoint field below.

Syncs can have multiple output models. This is useful when there is some form of dependency between the two models, e.g.

  • Fetching the data for a model from the external API requires fetching the data from another model first
  • You want the records of one model to be persisted before the records of another model

When defining multiple output models, you must define as many endpoints. The 1st endpoint will return the 1st model, the 2nd endpoint the 2nd model, etc.

Here’s an example:

  asana-fetch-tasks-and-comments:
    output:
      - AsanaTask
      - AsanaComment
    endpoint:
      - method: GET
        path: /tasks
      - method: GET
        path: /comments
endpoint
string|object
required

Defines the endpoint to use to fetch the output model.

endpoint:
  method: GET
  path: /tasks
  group?: Tasks

Possible method values are: GET

The method/endpoint combination can be shared across syncs to unify the communication with external APIs. The group is an optional value to group endpoints across scripts.

Legacy format: <METHOD> <URL-PATH>. e.g.: GET /tasks.

sync_type
incremental | full
required

Specifies whether each sync execution replaces all of the data (full) or upserts it (incremental).

Learn more about incremental vs. full refresh syncing.

runs
string

Specifies the frequency of the sync. Supports ms notations.

Defaults to 24h.

input
string

Defines the schema of the data required to run the sync. References a schema from the models.

Sync inputs are passed from your app to the sync script via the connection metadata (step-by-step guide).

scopes
string[]

Lists the necessary scopes to execute the sync.

This list is just indicative; it doesn’t trigger any automatic configuration (yet). List necessary scopes on the external API developer portal and your Nango UI integration settings.

Defaults to no scope.

auto_start
boolean

If true, automatically starts synchronizing between the external API and Nango when a new connection is created. Otherwise, it needs to be triggered via the API or Nango UI.

Defaults to true.

track_deletes
boolean

If true, automatically detects deleted records and flags them when you fetch the latest data.

Defaults to false.

This setting only applies if sync_type: full (details). For incremental syncs, this setting is ignored; instead, you must flag deleted records in the sync script using nango.batchDelete() (reference).

webhook-subscriptions
string | string[]

Lists the types of external webhooks the sync script will handle. Multiple syncs can listen to the same subscription. Learn more about handling external webhooks in syncs: step-by-step guide.

Action fields

Action configuration fields are under integrations.<INTEGRATION-ID>.actions.<ACTION-NAME>.

description
string
required

Describes the action.

output
string
required

Defines the schema of the data returned by the action. References a schema from the models section of this configuration file.

Each action has a corresponding generated endpoint to trigger it, configured in the endpoint field below.

endpoint
string|object
required

Defines the endpoint to use to trigger the action.

endpoint:
  method: POST
  path: /tasks

Possible method values are: GET, POST, PUT, PATCH, DELETE.

The method/endpoint combination can be shared across actions to unify the communication with external APIs. The group is an optional value to group endpoints across scripts.

Legacy format: <METHOD> <URL-PATH>. e.g.: POST /tasks.

input
string

Defines the schema of the data required to trigger the action. References a schema from the Models.

Action inputs are passed as parameters when triggering the action.

scopes
string[]

Lists the necessary scopes to trigger the action.

This list is just indicative; it doesn’t trigger any automatic configuration (yet). List necessary scopes on the external API developer portal and your Nango UI integration settings.

Defaults to no scope.

On-events fields

on-events configuration fields are under integrations.<INTEGRATION-ID>.on-events.<EVENT-NAME>.

Nango currently supports the following events:

  • post-connection-creation
  • pre-connection-deletion

Read the Execute scripts based on events guide.

post-connection-creation
string[]

Lists the scripts to execute after a new connection is created.

Example:

integrations:
  airtable:
    on-events:
      post-connection-creation:
        - setup # Name of your event-based script (without the .ts extension)
        - initialize-webhooks
pre-connection-deletion
string[]

Lists the scripts to execute before a connection is deleted.

Example:

integrations:
  airtable:
    on-events:
      pre-connection-deletion:
        - cleanup # Name of your event-based script (without the .ts extension)
        - remove-webhooks

Basic types

For each model, you must define its fields and data types. Data types are based on Typescript types, with some tweaks:

nango.yaml typeCorresponding Typescript type
bool or booleanboolean
stringstring
char or varcharstring
integer or int or number or bigint or floatnumber
dateDate
nullnull
undefinedundefined
anyany
arrayany[]
objectobject
true or falsetrue orfalse

Types not listed here are not supported. If you think we are missing something, please reach out in the Slack community.

Arrays

Your can use array types like this:

nango.yaml
models:
  Folder:
    files: string[]
    tickets: any[]

Type & string unions

You can use union types such as:

nango.yaml
models:
  User:
    name: string | null | undefined

You can also use string unions to enforce specific values:

nango.yaml
models:
  User:
    gender: m | f

You can mix types and strings in unions:

nango.yaml
models:
  User:
    gender: male | female | null

When you use the | notation, we check each element against known Typescript types, to categorize it as type (if match) or string (if no match).

Union array

Your can create a union array:

nango.yaml
models:
  Folder:
    issues:
      - PullRequest
      - Issues

Which translates to the following generated Typescript model:

export interface Task {
  gender: (PullRequest | Issues)[];
}

Reference other models

Your can use other models as types:

nango.yaml
models:
  User:
    id: string

  Account:
    users: User[]

Extend types

You can have a model extend the properties of another base model:

nango.yaml
models:
  Issue:
    id: integer
    title: string
    author: string

  GithubIssue:
    __extends: Issue
    issue_number: string

  LinearIssue:
    __extends: Issue
    roadmap_id: string

Dynamic field keys

You can allow dynamic keys in models:

nango.yaml
models:
  Task:
    __string: string
    id: string

Which translates to the following generated Typescript model:

export interface Task {
  [key: string]: string;
  id: string;
}

Optional property

You can set a property as optional:

nango.yaml
models:
  Task:
    name?: string

Unsupported syntax

A non-exhaustive list of unsupported syntax:

nango.yaml
models:
  Task:
    # Empty array
    name: []
    # Typescript utility types
    user: Pick<User>

Deploying your configuration

Changes to your integration configuration become active once you deploy them to a Nango environment using the nango deploy CLI command.

These changes can have significant impacts, in particular, if you:

  • Add a sync
  • Remove a sync or an action
  • Change a model or an integration script

Adding a sync

Unless you sync configuration specifies auto_start=false, the deployment will kick off the data synchronization between the external API and Nango, for all relevant connections.

This change will be reflected in the Nango UI in Integrations > select your integration > Endpoints tab.

Removing a sync

Removing a sync will erase all the cached data related to that sync, for all relevant connections. The synced records will no longer be accessive via the Nango API & SDKs.

This change will be reflected in the Nango UI in Integrations > select your integration > Endpoints tab.

Changing the integration ID in your configuration is the equivalent of deleting the sync with the old integration ID, and creating a new one. So, the data cached by the old sync will be deleted (for all relevant connections).

Removing an action

Removing an action will cause the failure of any request to trigger this action.

This change will be reflected in the Nango UI in Integrations > select your integration > Endpoints tab.

Changing a model or script

For full refresh syncs, as the data is replaced on each sync execution, the new model and script logic will apply to all records as soon as the sync executes following the deployment.

For incremental syncs, you can end up with model disparities, as older records will have outdated content. You will still be able to fetch them from the Nango cache. In some cases, it’s better to trigger a full resync via the API or the Nango UI, so that all historical records are fetched again using the new script & model.

Changing a model name in your configuration is the equivalent of deleting the model with the old name, and creating a new one.

Importing models in scripts

The models defined in your nango.yaml configuration are used to generate Typescript models (in the nango-integrations/models.ts file).

You can then import the relevant models from models.ts in your integration scripts.

If your app uses typescript, you can import these same models to enforce strong typing to the Nango API & SDK responses.

Questions, problems, feedback? Please reach out in the Slack community.