How to Build a Reliable Template Pivot

1. Filters (Preload Stage)

1. Select source boards in the Source Board stage.
2. Add preload filters to reduce fetch volume before joins.
3. Keep total preload filter fields at 4 or less.

2. Join Strategy (Before Transform)

MeshBoard supports multi-key joins across selected boards and lets you define explicit left/right column pairs for safer cross-board matching.

Supported join types

• Left: Keep all rows from your anchor board and match where possible.
• Inner: Keep only matched rows across both sides.
• Outer: Keep all rows from both sides (matched + unmatched).

Best practices

• Prefer multi-key joins for identity accuracy (for example, Email + Project ID).
• Pick one stable anchor board for left joins to avoid shifting totals run-to-run.

3. Column Mapping and Data Types

Map join keys, dimensions, and metrics explicitly. Before writeback, MeshBoard aligns output with destination monday column types to reduce GraphQL mutation failures.

• Use dimensions for grouping identity (owner, client, segment, region).
• Use metrics for calculations (sum/avg/min/max/count).
• Treat metric aliases as stable output contracts for dashboards and automations.

4. Transformation Options (Detailed)

Transformations run after joins and before final aggregation/output shaping. Use them to normalize keys, clean values, and reduce bad writeback outcomes.

4.1 String transforms (supported today)

• trim: remove leading/trailing whitespace.
• lower: lowercase normalization for case-insensitive keys.
• upper: uppercase normalization for code-like IDs.
• replace: literal substring replacement with pattern + replacement.
• titlecase: convert words to title case for display-friendly labels.

For replace, pattern is required and is matched as plain text (not regex). In the replacement field, leaving it empty removes the matched text; using " " (one space) replaces with whitespace.

4.2 Row filters (post-join filtering)

Use row filters to keep only meaningful records before aggregation.

• Text operators: equals, not_equals, contains, not_contains
• Numeric/date operators: gt (greater than), gte (greater than or equal), lt (less than), lte (less than or equal)
• Comparisons are case-insensitive for text matching.

4.3 Aggregation operations

• sum / avg: total values and average values.
• min / max: lowest and highest values.
• count: number of records in each group.

4.4 Date formatting and grouping

Date-family dimensions now support a dedicated Date Formatting manipulation with explicit grouping output.

• Shown only for date-family source columns: date, timeline, and week.
• Grouping options: Year, Year-Month, and Date.
• Default grouping is Date when not explicitly changed.
• If a selected dimension is not date-family, the Date Formatting option is hidden and the column stays as a normal dimension.

5. Transformation Guardrails by Plan

Transformation logic is available across plans, but execution scale and automation differ by tier.

• Basic: up to 3 boards, 5,000 rows per board, 250 output rows, manual runs only.
• Pro: up to 3 boards, 10,000 rows per board, 500 output rows, daily/weekly schedules.
• Ultimate: up to 5 boards, 15,000 rows per board, 500 output rows, hourly schedules.

Universal constraint: preload filters are capped at 4 fields on every tier.

6. Templates

A template is your saved Template Pivot blueprint. Save when you need repeatability across teams or reporting cycles.

• Basic: up to 1 saved pivot.
• Pro: up to 3 saved pivots.
• Ultimate: up to 5 saved pivots.

7. Automations

• Basic: manual trigger only.
• Pro: daily/weekly scheduling with hour selection.
• Ultimate: hourly sync enabled by system default.

If plan changes reduce capability, incompatible schedules are paused and can auto-resume after upgrade if they were paused due to downgrade.

8. Exports, Writeback, and Incremental Updates

• Use Preview before execute to open a popup that shows the first 5 aggregated rows.
• The preview popup opens immediately and shows a loading spinner while data is being prepared.
• Write aggregated rows to a summary board (existing or newly created).
• Export to CSV/JSON for external workflows.
• MeshBoard reads source data ordered by latest updates, then only creates or updates what changed in the target board for that Template Pivot.

In plain terms: your board stays fresh without you reloading everything from scratch every run. It behaves like an incremental data pipeline, but fully inside monday.

9. MCP Capabilities (Pro and Ultimate)

MCP helps you turn each Template Pivot result into next-step actions, so your team can run workflows in monday and beyond without rebuilding context each time.

• Use aggregated board output as the trusted trigger/input for external tool actions.
• Use the same output as a reliable context layer for monday AI agents, so agent decisions are based on cleaned and aggregated board data.
• Replicate monday-driven operational behavior in connected systems without manual copy-paste.
• Keep visibility in monday while execution can happen outside monday when needed.

If automated MCP provisioning is unavailable, use the permanent manual fallback: MCP Manual Setup and Recovery.