Claude’s Model Context Protocol (MCP) connects Claude to Morta so you can query data, perform controlled bulk updates, and evolve table schemas with clear, auditable prompts. Think of it as a high-leverage command line for your workspace: create reports based on your hub/s data, export snapshots, add/remove columns, and edit documents. Below are the core capabilities and bite-size use cases with short demos.
🧩 Capabilities at a glance
- Read: create reports and generate insights. Especially powerful across multiple hubs with standardized schemas.
- Write: insert/update rows (bulk operations)
- Schema: add columns, change types, create dependencies/lookups, create views.
⚠️ Safety & constraints (read this first)
- No Undo: Data changes and schema changes are permanent. You cannot undo changes except manually.
- Least privilege: Use the lowest role that can complete the task. Restrict the API key to a hub, document, or table/s
- Sandbox first: Validate in Sandbox; before applying on a live hub.
- Sensitive/Personal data: Data will be read out of Morta into Claude, hence it is important to avoid exposing sensitive/personal data.