Artie

Deletes a source reader unused by any active pipeline. Does not delete the source connector.

source_reader_delete

Deletes an SSH tunnel unused by any connector or pipeline.

ssh_tunnel_delete

Deletes an encryption key by UUID. Does not decrypt already-landed columns.

encryption_key_delete

Deletes an ingestion API key by UUID.

ingestion_api_key_delete

Permanently deletes a pipeline by UUID. Does not delete connectors.

pipeline_delete

Deploys a shared, validated source reader. To take a draft pipeline live, use pipeline_start.

source_reader_deploy

Checks the source for table or schema drift. To apply destination DDL, use pipeline_trigger_automatic_schema_changes.

pipeline_detect_schema_changes

Drops an inactive PostgreSQL replication slot. Check connector_fetch_postgres_replication_slots first: only an inactive or lost slot can be dropped.

connector_drop_postgres_replication_slot

Returns replication-relevant server settings (name, value, unit, category) for a saved Postgres source, such as wal_level, max_replication_slots, max_wal_senders, and max_slot_wal_keep_size. uuid must come from connector_list with includeSourceConnectors=true. Use to check CDC prerequisites (wal_level must be logical) before pipeline_start or after it fails. Settings are changed on the database, not through Artie.

connector_fetch_postgres_settings

Lists database names a saved source can see. uuid must come from connector_list with includeSourceConnectors=true. Do not pass a pipeline uuid or sourceReaderUUID.

connector_fetch_databases

Lists schemas on a saved connector. uuid must come from connector_list. Do not pass a pipeline uuid or sourceReaderUUID. Pass schemaName into connector_fetch_tables when the destination requires it.

connector_fetch_schemas

Returns columns and metadata for one table on a saved connector. uuid must come from connector_list (includeSourceConnectors=true for sources). Do not pass a pipeline uuid or sourceReaderUUID. tableName is required. This is column metadata, not pipeline lag or row counts.

connector_fetch_table_detail

Lists tables on a saved connector. uuid must come from connector_list. Do not pass a pipeline uuid or sourceReaderUUID. Use to choose names for pipeline_update.tables.

connector_fetch_tables

Returns SQL to create Oracle shadow tables for CDC. Oracle sources only.

connector_generate_shadow_script

Returns one PrivateLink connection by UUID.

private_link_connection_detail

Returns a saved connector by uuid from connector_list, not a pipeline uuid or sourceReaderUUID. Includes type, label, and sharedConfig. Sensitive fields may be masked placeholders. To ping that connector, pass this payload into unsaved_connector_ping and keep uuid set so stored secrets unmask.

connector_detail

Returns a source reader by uuid from pipeline_detail.sourceReaderUUID, not a pipeline uuid or connector uuid. Includes database, tablesConfig, and settings (replicationSlotOverride, publicationNameOverride, heartbeats). Sensitive fields may be present. Echo this into source_reader_update; do not invent a partial body. If pipeline_start 400s that two readers share slot artie, GET this, set settings.replicationSlotOverride to a unique lowercase name (digits and underscores, <=63), POST source_reader_update, then retry start.

source_reader_detail

Returns one SSH tunnel by UUID (no private key).

ssh_tunnel_detail

Returns FullPipeline for a uuid from pipeline_list: tables (uuid, name, schema, status), destinationUUID, specificDestCfg, and sourceReaderUUID. Echo this into pipeline_update; do not send pipeline_list. Use for per-table status and backfill table UUIDs. sourceReaderUUID is the uuid for source_reader_detail and source_reader_update, not a connector uuid. This is not lag (pipeline_usage) or error logs (pipeline_error_logs). Omit includeRelatedObjects — that Dashboard flag wraps connectors and is not this tool's success shape.

pipeline_detail

Returns per-table ingestion lag and row counts for a saved pipeline over a time range. Pass the pipeline uuid plus from and to as RFC3339 timestamps; set excludeBackfills=true to leave backfill rows out of row counts. Use when the user asks about lag, throughput, or rows processed. This is not slot health, warehouse correctness, pipeline_list status, pipeline_detail table status, or error logs (pipeline_error_logs).

pipeline_usage

Lists replication slots on a saved Postgres source (name, database, active, walStatus, retainedWALBytes, isReplica). uuid must come from connector_list with includeSourceConnectors=true, not a pipeline uuid or sourceReaderUUID; pass databaseName for the pipeline's reader database. Use for slot health and WAL retention, and to confirm a slot is inactive before connector_drop_postgres_replication_slot.

connector_fetch_postgres_replication_slots

Lists column hashing salt names and UUIDs (no secret).

column_hashing_salt_list

Lists saved connectors. Defaults to destinations; pass includeSourceConnectors=true for sources. Call this before connector_fetch_databases, connector_fetch_schemas, connector_fetch_tables, or connector_fetch_table_detail — those tools need a uuid from this list, not a pipeline uuid or sourceReaderUUID from pipeline_list.

connector_list

Lists encryption key names and UUIDs (no secret material).

encryption_key_list

Returns recent error logs for a saved pipeline (message, errorDetail, table/schema, timestamps). Pass the pipeline uuid from pipeline_list. Use when the user asks why a pipeline failed or is unhealthy. This is not lag (pipeline_usage), per-table status (pipeline_detail), or replication slot size.

pipeline_error_logs

Lists pipelines with uuid, name, and status. This is not a full configuration — do not send it to pipeline_update. Full config, per-table status, and table UUIDs are pipeline_detail. Failure messages are pipeline_error_logs, not status. sourceReaderUUID is not a connector id; call connector_list before connector_fetch_table_detail or other connector_fetch_* tools.

pipeline_list

Pings a connector before connector_create. Pass type, sharedConfig (host, user, password, and type-specific fields), and connectionRole (source or destination). This is not a uuid ping — do not pass a pipeline uuid or sourceReaderUUID. Postgres needs a database name, Oracle a service name, S3 a bucket. Success is {"success": true}; a JSON error field means the host or credentials failed.

unsaved_connector_ping

Hashes a sample string with this salt and returns the hex digest.

column_hashing_salt_preview

Searches ingested catalog metadata. For live source tables, use connector_fetch_tables.

data_catalog_search

Starts a DynamoDB point-in-time export. DynamoDB sources only; for SQL CDC use pipeline_start.

connector_start_dynamo_db_export

Deploys a draft or resumes a paused pipeline once destination, tables, and (Postgres/Cockroach/Oracle) reader database are set. Do not call immediately after create-from-source. If start 400s that two source readers share replication slot artie, GET source_reader_detail for sourceReaderUUID from pipeline_detail, set settings.replicationSlotOverride to a unique name, POST source_reader_update with the echoed reader, then retry start.

pipeline_start

Enqueues automatic schema changes for all eligible running pipelines. For one pipeline, use pipeline_trigger_automatic_schema_changes.

company_trigger_automatic_schema_changes

Applies detected source DDL on one pipeline. For every eligible pipeline, use company_trigger_automatic_schema_changes.

pipeline_trigger_automatic_schema_changes

Updates a PrivateLink connection. Region, VPC service name, data plane, and AZs are only mutable while under review or errored.

private_link_connection_update

Updates the name and description of an existing column hashing salt. Does not rotate the secret.

column_hashing_salt_update

Applies a SourceReader from source_reader_detail. Echo the last GET and change only the intended fields. Omitting name, database, or connectorUUID blanks them. To fix a duplicate Postgres slot (pipeline_start 400 that two readers share artie), set settings.replicationSlotOverride to a unique lowercase name, then pipeline_start. This is not pipeline_update or connector_update.

source_reader_update

Updates name, host, port, or username on an existing SSH tunnel. Does not rotate keys.

ssh_tunnel_update

Updates the name and description of an existing encryption key. Does not rotate the key material.

encryption_key_update

Updates metadata on an existing ingestion API key. Does not mint a new key.

ingestion_api_key_update

Replaces the full pipeline configuration (not a PATCH). Echo the last FullPipeline from pipeline_create_from_source, pipeline_update, or pipeline_detail — not pipeline_list — and send destinationUUID and tables together; a destination-only body is rejected and omitting tables deletes all tables.

pipeline_update