Relations
Relations are the core primitive. A database without relations is a spreadsheet; with them, your databases become a connected model of your business.
Paired on both sides
Section titled “Paired on both sides”A relation is a first-class object that ties two databases together. Creating one always creates two relation-fields — one on each database — pointing at the same relation. Link Projects and Tasks once, and:
- a Project record shows its Tasks, and
- a Task record shows its Project.
Both directions are navigable by construction — there are no orphaned one-way links.
Cardinality
Section titled “Cardinality”- One-to-many — e.g. one Project has many Tasks; each Task has one Project.
- Many-to-many — e.g. Articles ↔ Authors.
Self-relations (a database related to itself, e.g. Task → parent Task) and cross-space relations are both allowed.
Working with relations
Section titled “Working with relations”- In the UI, the relation picker is first-class — pick target records by title.
- Via the API, relation values are not part of a record’s
values; they live in a separate links table and are returned as{id, title}chips. Set them with the links endpoints or the MCPlink_recordstool. See querying records.
What relations unlock
Section titled “What relations unlock”Once two databases are related, you can:
- Look up a related record’s field (surface the Client’s owner on each Project).
- Roll up related records (sum a Client’s Opportunity amounts, count open Tasks).
- Reference both in formulas (
{Budget} - {Spent}).