Data Connectors
Using MCPs with Data Connector
If you’re exploring how to use MCPs (Modular Command Pipelines) with our new Data Connector feature, you’re in the right place. This guide gives you a quick overview of what MCPs are, why they matter, and how you can use them to power advanced workflows in Julius.
What Are MCPs?
MCPs (short for Modular Command Pipelines) are a powerful way to define structured, repeatable operations within Julius. Think of them as reusable blocks of logic that you can apply to data- kind of like a saved prompt, but with clearer parameters and composability. Each MCP consists of:- Inputs: Parameters the user or system supplies (e.g., a data source, a date range, a target column).
- Logic: A templated instruction or chain of instructions Julius will follow.
- Output: Structured or freeform results, such as cleaned tables, summaries, or decisions.
How MCPs + Data Connector Work Together
Here’s how you can use them together:- Connect Your Data
- Go to your Julius workspace settings and link your data source via the Data Connector.
- Run the MCP on Connected Data
- Select your connected data source in the MCP interface.
- Provide any required inputs (column names, filters, thresholds).
- Click Run—Julius will apply the MCP logic to your live data.
- Review, Iterate, Export
- Julius returns results in the format you define.
- You can adjust inputs or refine the logic as needed.
- Export to CSV or sync back to your data source.
Supported MCPs (more to come!):
| Notion | GitHub |
|---|---|
| Linear | RenvenueCat |
| Stripe | Intercom |
| BrowserBase |