Connecting Your Own Data

Lomita works with any data source — not just the 104 in the catalog. Connect your CRM, ERP, internal APIs, databases, or upload files. AI agents handle the integration automatically.

From the chat

Tell the chat what you want to connect:

Connect my HubSpot — I need deal pipeline and close dates

Connect our internal sales API at api.company.com

The Discovery Agent will ask for:

API documentation URL (helps the agent understand endpoints)
Authentication type (API key, OAuth, bearer token)
Credentials (encrypted with AES-256-GCM, stored securely)
What specific data you need

The Integration Engineer then reads the docs, builds a Dagster data pipeline, and starts ingesting data — all automatically.

From your AI agent

connect_integration(
  service: "hubspot",
  credentials: { api_key: "pat-na1-..." },
  config: {
    goal: "Pull deal pipeline data — amounts, stages, close dates",
    docs_url: "https://developers.hubspot.com/docs/api"
  }
)

From the Sources page

Click Connect via Form on the Sources page to fill in:

Service name
API credentials (key-value pairs)
What data you need (plain English)
API documentation URL (optional, helps the agent)
Sync schedule (hourly, daily, weekly)

What happens next

Credentials are encrypted and stored
The Integration Engineer receives a task
The agent reads the API documentation
A Dagster data pipeline is created and deployed
Data starts flowing into your private data lake
The source becomes available for all future research

Uploading files

For one-off data (CSV, JSON, Parquet):

upload(filename: "sales-q1.csv", content: "...")

Or use the Sources page upload feature. Uploaded data is queryable immediately via the analytics engine.

Security

Credentials encrypted with AES-256-GCM at rest
Decrypted only at pipeline runtime by the Integration Engineer
Stored in your tenant's isolated PostgreSQL database
Never shared across tenants
API key rotation supported via the Sources page

Lomita Documentation