Create column

Path parameters

• table_idrequired

  string

  Target table ID.

Body parameters

• keyrequired

  string

  Stable machine key for the column.

• labelrequired

  string

  Human-readable column label.

• kind · optional

  string

  Column kind: manual, source, formatter, enrichment, json-explode, projection, external_id, or system.

• dataType · optional

  string

  Data type: text, number, date, boolean, email, url, json, or message (default text).

• textFormat · optional

  string

  Text rendering format: plain, markdown, long_text, or code.

• origin · optional

  string

  Column origin: manual, csv, api, ai, email_finder, email_validator, phone_finder, outreach, onboarding, crm_export, list_builder, agent_run, agent_source, or system.

• source · optional

  object

  Optional source metadata object.

• sourceId · optional

  string

  Optional source identifier.

• importVersion · optional

  number

  Optional import version metadata.

• fieldSemantics · optional

  object

  Optional field semantics metadata.

• valueSchema · optional

  object

  Optional value schema metadata.

• version · optional

  number

  Optional column version metadata.

• status · optional

  string

  Column status: idle, running, or error.

• costEstimate · optional

  number

  Optional cost estimate metadata.

• lastRunAt · optional

  string

  Optional last-run timestamp metadata.

• errorMessage · optional

  string

  Optional last error message metadata.

• isHidden · optional

  boolean

  Whether the column is hidden.

• position · optional

  number

  Optional explicit float sort order. If provided, afterColumnId/beforeColumnId are ignored.

• afterColumnId · optional

  string

  Optional insertion anchor. Place the new column after this column when position is omitted.

• beforeColumnId · optional

  string

  Optional insertion anchor. Place the new column before this column when position is omitted.

• suppressSaveAutoRun · optional

  boolean

  Save-hook control (default false on create). When true, suppresses save-time auto-run scheduling; ordinary clients should omit it.

• isImportant · optional

  boolean

  Marks the column as a favorite/starred context field for draft and calling fallback.

• config · optional

  object

  Kind-specific configuration payload.

• config.provider · optional

  string

  Optional hint. Server policy decides final provider/model.

• config.selectedRowFields · optional

  array<string>

  For LLM enrichment: exact source column keys (or key.path tokens) to include in each row input. Select keys available in the table schema; this is input selection, not a workspace catalog field ID.

• config.selectedRowFieldRefs · optional

  array<object>

  Optional stable source bindings for selectedRowFields. Each reference has snake_case column_id, column_key, and path strings; column_id must identify the selected source column, and path may be empty for the whole value.

• config.selectedRowFieldRefs[].column_id · optional

  string

  Stable source-column ID. References missing a complete binding are ignored.

• config.selectedRowFieldRefs[].column_key · optional

  string

  Current source column key; together with path it must form a selectedRowFields token.

• config.selectedRowFieldRefs[].path · optional

  string

  Optional nested JSON path; omitted or empty selects the whole column value.

• config.destinationTableId · optional

  string

  For sync_to_table single-destination mode.

• config.columnMappings · optional

  array<object>

  For sync_to_table single-destination mappings (sourceKey/destKey) or route-level mappings.

• config.routes · optional

  array<object>

  For sync_to_table router mode (evaluated top-to-bottom); each route can define condition, syncMode, syncBehavior, listSourceColumnId, listPath, and columnMappings.

• config.syncMode · optional

  string

  For sync_to_table mode: row or list.

• config.syncBehavior · optional

  string

  For sync_to_table behavior: link or copy.

• config.listSourceColumnId · optional

  string

  For sync_to_table list mode: source JSON column ID whose canonical items[] payload should be expanded.

• config.listPath · optional

  string

  For sync_to_table list mode: optional item array path (default items).

• config.columnMappings[].sourceScope · optional

  string

  For sync_to_table list mode: row or item.

• config.columnMappings[].itemKey · optional

  string

  For sync_to_table list mode when sourceScope=item: item field name to map.

• config.columnMappings[].sourceKey · optional

  string

  For sync_to_table when sourceScope=row: parent/source row column key to map.

• config.columnMappings[].path · optional

  string

  For sync_to_table when sourceScope=row: optional nested JSON path inside sourceKey.

• config.fieldMappings · optional

  object

  For add_to_crm / Add to Leads: field mapping contract from source columns into canonical lead fields.

• config.fieldMappings.companyDomain · optional

  string

  For add_to_crm: source column key for the company domain. Required.

• config.fieldMappings.linkedinUrl · optional

  string

  For add_to_crm: source column key for LinkedIn URL (optional but recommended).

• config.fieldMappings.emailAddresses · optional

  array<object>

  For add_to_crm: email mappings. Each mapping can point to a primitive column or a structured source column plus explicit path.

• config.fieldMappings.emailAddresses[].column · optional

  string

  For add_to_crm email mappings: source column key.

• config.fieldMappings.emailAddresses[].path · optional

  string

  For add_to_crm email mappings: required when the source column stores JSON/object output (example: response).

• config.fieldMappings.phoneNumbers · optional

  array<object>

  For add_to_crm: phone mappings. Each mapping can point to a primitive column or a structured source column plus explicit path.

• config.fieldMappings.phoneNumbers[].column · optional

  string

  For add_to_crm phone mappings: source column key.

• config.fieldMappings.phoneNumbers[].path · optional

  string

  For add_to_crm phone mappings: required when the source column stores JSON/object output (example: mobile_number).

• config.fieldMappings.common · optional

  object

  For Add to Leads multi-source mode: parent-row mappings applied to every lead source, usually companyDomain and companyName.

• config.fieldMappings.leadSources · optional

  array<object>

  For Add to Leads multi-source mode: row and array lead sources to materialize from one table row.

• config.fieldMappings.leadSources[].kind · optional

  string

  For Add to Leads multi-source mode: row or array.

• config.fieldMappings.leadSources[].sourceColumn · optional

  string

  For Add to Leads array source: source JSON column key.

• config.fieldMappings.leadSources[].sourceArrayPath · optional

  string

  For Add to Leads array source: array path inside sourceColumn, commonly items.

• config.fieldMappings.leadSources[].template · optional

  object

  For Add to Leads multi-source mode: canonical lead fields mapped from row columns or __array_item paths.

• config.sourceColumns · optional

  array<string>

  For add_to_crm: optional explicit dependency/source column list.

• config.sequenceId · optional

  string

  For add_to_sequence: target sequence id. The sequence must be ACTIVE and its sourceTableId must match the table this column belongs to.

• config.sourceLeadColumn · optional

  string

  For add_to_sequence: the column key (not field name) of an add_to_crm or lead_finder column whose cell output contains lead IDs.

• config.assignedToUserId · optional

  string

  For add_to_sequence: assignee for generated sequence tasks. Easy path: set this and omit senderAccountIds/assignmentPlan.

• config.skipActiveSequenceConflicts · optional

  boolean

  For add_to_sequence: skip people already active in another sequence. Recommended for automated/table-driven handoffs.

• config.senderAccountIds · optional

  object

  For add_to_sequence advanced routing: explicit sender account IDs keyed by email/linkedin.

• config.assignmentPlan · optional

  array<object>

  For add_to_sequence advanced/team routing: routes with assigneeUserId, optional senderAccountIds, and optional weight.

• config.routingMode · optional

  string

  For add_to_sequence advanced/team routing: single, round_robin, or weighted_round_robin.

• config.url · optional

  string

  For http_request: URL template. Use {{column_key}} for row values and {{secrets.name}} for workspace secrets.

• config.method · optional

  string

  For http_request: HTTP method (GET, POST, PUT, PATCH, DELETE, HEAD, OPTIONS).

• config.headers · optional

  object

  For http_request: key-value header pairs. Values support {{secrets.name}} templates.

• config.body · optional

  string | object

  For http_request: request body template. Can be null, a string template, or an object sent as JSON.

• config.timeoutSeconds · optional

  number

  For http_request: request timeout in seconds (default 30).

• autoRun · optional

  string

  Auto-run trigger: never, onInsert, or onSourceUpdate (default never).

• config.runCondition · optional

  RunCondition

  Optional recursive column-owned predicate. Applied to candidate rows for auto-run, /run, and /estimate; omit to consider all candidate rows.

Definitions

{
  "type": "all",
  "children": [
    {
      "type": "condition",
      "columnKey": "linkedin_url",
      "operator": "isNotEmpty",
      "missingPolicy": "unknown"
    },
    {
      "type": "condition",
      "columnKey": "country",
      "operator": "equals",
      "missingPolicy": "unknown",
      "value": "United States"
    }
  ]
}

Errors