Data Connectors
Bring your own data sources into Jetro.
Built-in data
Out of the box, Jetro pulls financial data through its own API. That covers most publicly traded equities. But your best data probably lives somewhere else.
Adding your own sources
Just tell your agent what you want to connect to. Say something like “connect to my Postgres database” or “pull data from the Stripe API” and the agent builds the connector for you. No forms to fill out, no config files to write.
If it has an API or a database connection string, you can connect it. The agent handles the setup, auth, and schema mapping.
Databases
Common database types the agent can connect to include PostgreSQL, MySQL, SQLite, DuckDB, MotherDuck, Amazon S3, Snowflake, and BigQuery. But anything with a standard connection protocol works.
APIs
Any REST API is fair game. Bloomberg, Morningstar, FactSet, your own internal services, public government data, anything. Tell the agent what API you want, give it your credentials, and it wires up the connector.
API keys and passwords are stored in your OS keychain via VS Code SecretStorage. They never end up in plain text config files or your project directory.
Managing connectors
The Connector Panel gives you a visual overview of all your connections. Browse table schemas, test that a connection works, and import tables into your local DuckDB cache for fast querying.
Universal vs project-scoped
Add a universal connector from the Library tab to make it available across all projects. Or add a project-scoped connector from within a specific project to keep it isolated.
Once a connector exists, the agent can query it like any other data source. No extra setup needed per canvas or per prompt.
Two-way connectors
Connectors aren’t just for pulling data in. They work in both directions. The agent can push data out too. Post a summary to Slack, update a Google Sheet, publish a chart to your WordPress blog, write results back to a database. If the service has an API, the agent can read from it and write to it.