Data Source Integration

Outter is designed to integrate with external data sources so that your AI-powered features can leverage your existing data. This is crucial for SMEs who have data in cloud databases or third-party services. You can connect Outter to various cloud-hosted databases like AWS RDS (MySQL/PostgreSQL), Azure SQL Database, Firebase/Firestore, or other sources. The platform provides API endpoints to register, validate, and fetch data from these external sources, all under a secure and unified interface.

Connecting External Databases

To connect an external database, you will typically register a data source via an API call. For example, you might use an endpoint like POST /api/v2/integrations/data-sources with a JSON payload containing the connection details (type of database, connection URI or credentials, etc.). Outter’s Integration Connector Service will store these credentials securely (encrypted, often in a secrets vault). Once registered, Outter can either pull data on demand or on a schedule, or accept push updates from your system:

• Pull Model (Outter fetches your data): After registering, Outter can periodically poll the external system’s API or database to retrieve new data. For instance, you might configure Outter to poll an AWS RDS instance every hour for new entries in a customer_feedback table. The connector service would use the stored credentials to connect to your database (over SSL) and fetch the data securely. Each polling interval or fetch action can be triggered via an endpoint (e.g., POST /api/v2/integrations/data-sources/{id}/poll) or automatically by Outter's scheduler based on your configuration.
• Push Model (external system pushes to Outter): Alternatively, your external system can call Outter whenever new data is available. In this case, Outter provides a webhook-like endpoint for your data source. For example, a CRM could send an HTTP POST to https://api.outter.io/v2/integrations/crm_hook with a payload whenever a new note is created, and Outter will ingest and process that. These custom integration endpoints are secured with their own API key or token to ensure only your system can push data.

Validation: Outter provides a way to validate the connection to your data source. After registering a data source, you can call an endpoint like POST /api/v2/integrations/data-sources/{id}/validate to have Outter attempt a test connection. The response will confirm a successful connection or provide error details (e.g., invalid credentials or network issues). This helps ensure your configuration is correct before relying on data in production.

Fetching Data in AI Requests: Once connected, you can reference these data sources in your AI API requests if needed. For example, when calling a recommendations endpoint, you might specify to use a particular data source for user purchase history. Or Outter might automatically use the integrated data (for example, the Recommendation service could use product embeddings or transaction history from your database, which Outter keeps synchronized). Typically, you'd ensure the relevant data is either fetched ahead of the AI call or included in the request context.