2026-09-28AutoDataOpen in dashboard

Connector Reference

Every connection field, default and capability for all 38 AutoData connectors.

Field-by-field reference for all 38 connector types. Keys are the names used in saved connections and in the API’s secrets / inline_secrets. “*” marks fields required only when no full connection string is given; “API only” keys are accepted by the server but not offered in the dashboard forms. For how to connect, see Connector Data Sources. Every type except Google Analytics 4 can also be a write destination; the modes each one offers are listed here, and the details are in Write Output & Auto-Sink.

Databases

SQL Database — `sql`

Tables and views of any SQLAlchemy-reachable database (PostgreSQL, MySQL/MariaDB, SQL Server, SQLite).

Key	Required	Default	Notes
`connection_string`	Yes	—	SQLAlchemy URL, e.g. postgresql://user:pass@host:5432/db?sslmode=require, mysql+pymysql://…, mssql+pyodbc://…?driver=ODBC+Driver+18+for+SQL+Server. Percent-encode special characters in the password.
`ssl_root_cert_pem`	No	—	PostgreSQL: server CA certificate, pasted as PEM text.
`ssl_cert_pem / ssl_key_pem`	No	—	PostgreSQL: client certificate and key, pasted as PEM text.
schema (API only)	No	database default	Schema that Discover lists.

Custom query: yes — the only connector with the Write Query box in the dashboard.
Incremental: yes.
Write destination: yes (replace, append, fail, merge).

MongoDB — `mongodb`

Documents of one collection; nested documents become dotted columns; _id is dropped. Table = database.collection.

Key	Required	Default	Notes
`connection_string`	Yes	—	mongodb+srv://… (Atlas) or mongodb://…; the scheme decides TLS.
database (API only)	No	autodata	Database for Discover and for bare collection names.
tls (API only)	No	from URI; runs: true	false for plaintext servers.
tls_allow_invalid_certificates (API only)	No	false	Accept self-signed certificates.

Custom query: no.
Write destination: yes (database.collection; replace, append, fail, merge).
Known issue (2026-09-28): runs that read from MongoDB currently fail with a server error. Test, Discover, Preview and writing to MongoDB work.

Oracle Database — `oracle`

Tables and views (oracledb thin driver, no Instant Client).

Key	Required	Default	Notes
`connection_string`	Or split fields	—	oracle+oracledb://user:pwd@host:1521/?service_name=ORCLPDB1
`host`	Yes*	—
`port`	No	1521
`service_name / sid`	One of them*	—
`user`	Yes*	—
`password`	Yes*	—
`schema`	No	user’s schema	Owner Discover lists.

* unless connection_string is given.
Custom query: Oracle SQL (API).
Incremental: yes. Write destination: yes (replace, append, fail, merge).

Apache Cassandra — `cassandra`

Rows of a table in a keyspace (Cassandra, DSE, ScyllaDB).

Key	Required	Default	Notes
`contact_points`	Yes	—	host1,host2
`port`	No	9042
`username / password`	No	—
`keyspace`	For reads	—
`datacenter`	No	—	Local DC.
allow_filtering (API only)	No	false	Needed for incremental reads on non-clustering columns.

Custom query: no. Incremental: yes. Write destination: yes (append, replace, fail; merge_keys name the primary key of a new table).

Amazon DynamoDB — `dynamodb`

All items of a table (Scan).

Key	Required	Default	Notes
`access_key_id`	Yes	—
`secret_access_key`	Yes	—
`session_token`	No	—	For temporary keys.
`role_arn / external_id`	No	—	A role to assume from the access keys.
`region`	Yes	—
`endpoint_url`	No	—	DynamoDB Local / VPC endpoint.

Preview row count is DynamoDB’s approximate ItemCount.
Incremental: yes (scan filter). Write destination: yes (append, fail; merge_keys name the key of a new table).

Elasticsearch / OpenSearch — `elasticsearch`

Documents of an index; nested fields become dotted columns.

Key	Required	Default	Notes
`url`	Yes (or cloud_id)	—
`cloud_id`	Instead of url	—	Elastic Cloud deployment ID.
`api_key`	No	—	Used first, then the bearer token, then user and password.
`bearer_token`	No	—
`username / password`	No	—
`verify_certs`	No	true	false for self-signed dev clusters.
`ca_cert_pem`	No	—	CA certificate of a private CA, pasted as PEM text.

Discover hides dot-prefixed system indices.
Incremental: yes (range query). Write destination: yes (index; replace, append, fail, merge).

Data warehouses

Snowflake — `snowflake`

Tables in one database; table = TABLE or SCHEMA.TABLE. Sign in with a password, a key pair or a token; an optional role.

Key	Required	Default	Notes
`account`	Yes	—	orgname-accountname (no .snowflakecomputing.com).
`user`	Yes	—	`username` is accepted in its place.
`password`	One sign-in method	—
`private_key_pem`	One sign-in method	—	Key pair: the private key, pasted as PEM text.
`private_key_passphrase`	No	—	For an encrypted private key.
`token`	One sign-in method	—	An access token; `authenticator` defaults to oauth.
`database`	Yes	—	As Snowflake shows it (usually UPPERCASE).
`schema`	No	PUBLIC
`warehouse`	No	user default
`role`	No	user default

Custom query: Snowflake SQL (API). Incremental: yes.
Write destination: yes (table upper-cased; replace, append, fail, merge).

BigQuery — `bigquery`

Tables as dataset.table or project.dataset.table.

Key	Required	Default	Notes
`project_id`	Yes	—	Project that runs the queries.
`service_account_json`	Yes	—	Full JSON key (alias credentials_json); needs BigQuery Data Viewer + Job User.

Custom query: GoogleSQL (API). Incremental: yes.
Write destination: yes (dataset.table; replace, append, fail, merge).

Amazon Redshift — `redshift`

Tables and views over the PostgreSQL protocol.

Key	Required	Default	Notes
`connection_string`	Or split fields	—
`host`	Yes*	—
`port`	No	5439
`database`	Yes*	—
`user / password`	Yes*	—
`schema`	No	public
`ssl_mode`	No	require
`ssl_root_cert_pem`	No	—	Server CA certificate, pasted as PEM text.
`ssl_cert_pem / ssl_key_pem`	No	—	Client certificate and key, pasted as PEM text.

Custom query: yes (API). Incremental: yes. Write destination: yes (replace, append, fail, merge; up to 1,000,000 rows per write).

ClickHouse — `clickhouse`

Tables and views over HTTP (default) or native protocol.

Key	Required	Default	Notes
`connection_string`	Or host	—
`host`	Yes*	—
`port`	No	8443 / 8123	TLS / plain
`database`	No	default
`user`	No	default
`password`	No	empty
`secure`	No	true

Custom query: yes (API). Incremental: yes. Write destination: yes (replace, append, fail; option order_by).

Databricks — `databricks`

Tables via a SQL warehouse; table = [catalog.]schema.table.

Key	Required	Default	Notes
`host`	Yes	—	Workspace hostname.
`token`	One sign-in method	—	Personal access token (dapi…).
`client_id / client_secret`	One sign-in method	—	Service principal.
`tenant_id`	No	—	Azure Databricks with an Entra ID service principal.
`http_path`	Yes	—	/sql/1.0/warehouses/…
`catalog / schema`	No	—	Discover scope; write target (defaults main / default).

Test times out after 10 s — start the warehouse first.
Custom query: yes (API). Incremental: yes.
Write destination: yes (replace, append, fail, merge).

Microsoft Fabric — `fabric`

Warehouse / Lakehouse SQL endpoint tables, as an Entra ID service principal.

Key	Required	Default	Notes
`workspace_id`	Yes	—	The SQL endpoint host prefix (<id>.datawarehouse.fabric.microsoft.com).
`tenant_id`	Yes	—
`client_id`	Yes	—
`client_secret`	Yes	—	Secret value.

The service principal must be a workspace member and the tenant must allow service principals.
Custom query: T-SQL (API). Incremental: yes.
Write destination: yes (replace, append, fail, merge).

Azure Synapse — `synapse`

Dedicated SQL pool tables and views.

Key	Required	Default	Notes
`workspace`	Yes*	—	Name or full server.
`database`	Yes*	—	SQL pool.
`port`	No	1433
`user / password`	SQL auth*	—
`client_id / client_secret / tenant_id`	AAD auth*	—
`schema`	No	dbo

Custom query: T-SQL (API). Incremental: yes. Write destination: yes (replace, append, fail, merge).

SAP HANA — `sap_hana`

Tables and views (HANA Cloud or on-premise).

Key	Required	Default	Notes
`connection_string`	Or split fields	—
`host`	Yes*	—
`port`	No	443	On-prem 3<nn>15.
`user / password`	Yes*	—
`schema`	No	user’s
`encrypt`	No	true

Custom query: yes (API). Incremental: yes. Write destination: yes (replace, append, fail, merge).

Storage and files

Amazon S3 — `s3`

One object per run (CSV, TSV/TXT, Parquet, JSON, JSON Lines, Excel, Feather, ORC). Table = object key.

Key	Required	Default	Notes
`bucket`	Yes	—
`region`	No	us-east-1
`access_key_id`	Yes	—
`secret_access_key`	Yes	—
`session_token`	No	—	For temporary keys.
`role_arn / external_id`	No	—	A role to assume from the access keys.
`prefix`	No	—	Folder for Discover.
`endpoint_url`	No	Amazon S3	An S3-compatible store.

Test checks the connection’s bucket (needs s3:ListBucket on it).
Custom query: S3 Select on CSV (API).
Write destination: yes (a file under the prefix; replace, append, fail; file_format). Listener source.
Known issue: with a prefix set, a file picked from Discover fails at run time because the prefix is added twice. Leave Prefix empty on connections you read from.

Google Cloud Storage — `gcs`

One object per run. Table = full object name.

Key	Required	Default	Notes
`project_id`	No	key’s project
`bucket`	Yes	—
`service_account_json`	Yes	—	Storage Object Viewer on the bucket.

Write destination: yes (a file under the prefix; replace, append, fail; file_format). Listener source.

Azure Blob Storage — `azure_blob`

One blob per run.

Key	Required	Default	Notes
`connection_string`	One auth option	—
`account_url`	One auth option	—
`account_key / sas_token`	With account_url	—
`tenant_id / client_id / client_secret`	With account_url	—	Service principal.
`auth_method`	No	first method stored	connection_string, sas, account_key or service_principal.
`container`	For reads	—
`prefix`	No	—	See the prefix issue under S3; the same applies.

Write destination: yes (a file under the prefix; replace, append, fail; file_format). Listener source.

Azure Data Lake Storage Gen2 — `adls_gen2`

One file per run from a filesystem.

Key	Required	Default	Notes
`connection_string`	One auth option	—
`account_name`	One auth option	—
`account_key / sas_token`	With account_name	—
`tenant_id / client_id / client_secret`	With account_name	—	Service principal.
`auth_method`	No	first method stored	connection_string, sas, account_key or service_principal.
`filesystem`	For reads	—
`path`	No	—	Same double-prefix issue as S3.

Write destination: yes (a file under the path; replace, append, fail; file_format). Listener source.

Delta Lake — `delta`

The latest version of one Delta table (no Spark).

Key	Required	Default	Notes
`path`	Yes	—	s3://…, abfss://…, gs://… or a server path.
`access_key_id / secret_access_key / region / endpoint_url / session_token`	No	—	S3 and S3-compatible.
`role_arn / external_id`	No	—	S3: a role to assume from the access keys.
`storage_options`	No	—	JSON of delta-rs options, e.g. Azure or GCS keys.

The whole table is loaded before the row cap.
Write destination: yes (replace, append, fail, merge).

Google Sheets — `google_sheets`

One worksheet; public sheets need no credentials.

Key	Required	Default	Notes
`url`	Yes	—	Spreadsheet URL or ID.
`gid`	No	0	Tab id for public sheets.
`service_account_json`	Private sheets	—	Share the sheet with the service account.

Write destination: yes (worksheet; replace, append, fail, merge), with a service account that can edit the sheet.

Google Drive — `google_drive`

One file (native Sheets exported as CSV).

Key	Required	Default	Notes
`service_account_json`	Yes	—	Share files with its email.
`file_id`	No	—	ID or share URL.
`folder_id`	No	—	Discover scope.
`delegated_user`	No	—	Domain-wide delegation.

Write destination: yes (a file in a folder; option folder_id; replace, append, fail; file_format).

Dropbox — `dropbox`

One file; table = /path/file.csv.

Key	Required	Default	Notes
`access_token`	One option	—
refresh_token + app_key + app_secret	One option	—	Auto-renewing.
`path`	No	root	Folder for Discover or file to read.

Write destination: yes (a file in the connection’s folder; replace, append, fail; file_format).

OneDrive / SharePoint Files — `onedrive`

One file in a drive or document library (Microsoft Graph, app-only).

Key	Required	Default	Notes
`tenant_id / client_id / client_secret`	Yes	—	Files.Read.All or Sites.Read.All (application).
`user_id / drive_id / site_url`	One of them	—	Target drive.
`path`	No	—	File path.

Write destination: yes (a file in the connection’s drive; replace, append, fail; file_format).

Box — `box`

One file; table = file id or “name (id)”.

Key	Required	Default	Notes
`jwt_config_json`	One option	—	JWT app config.
`developer_token`	One option	—	Expires in 60 minutes.
`file_id`	No	—
`folder_id`	No	0	Discover scope.

Write destination: yes (a file in a folder; option folder_id; replace, append, fail; file_format).

Streaming, IoT and historians

Apache Kafka — `kafka`

A bounded batch from one topic; JSON objects become rows.

Key	Required	Default	Notes
`bootstrap_servers`	Yes	—
`topic`	No	—	Overrides the Discover pick.
`group_id`	No	autodata-consumer	Committed offsets decide where the next read starts.
`security_protocol`	No	from the other fields	PLAINTEXT, SSL, SASL_PLAINTEXT, SASL_SSL.
`sasl_mechanism`	SASL	—	PLAIN, SCRAM-SHA-256/512.
`sasl_plain_username / sasl_plain_password`	SASL	—
`ca_cert_pem`	No	—	CA certificate of a private CA, pasted as PEM text.
`client_cert_pem / client_key_pem`	No	—	Client certificate and key, pasted as PEM text.

Test, Discover, runs and streams sign in the same way: SASL, a custom CA and a client certificate as stored.
Preview returns fixed columns and the partition count, not message fields; start runs via the API and name the target field.
Stops after 10 s without messages. Write destination: yes (topic; append; option partition_key_column). Streaming source/sink (Enterprise).

Amazon Kinesis — `kinesis`

A bounded batch from all shards, from the oldest retained record every time.

Key	Required	Default	Notes
`stream_name`	Yes	—
`region`	No	us-east-1
`access_key_id / secret_access_key`	Yes	—
`session_token`	No	—	For temporary keys.
`role_arn / external_id`	No	—	A role to assume from the access keys.
`endpoint_url`	No	Amazon Kinesis	A custom endpoint.

Preview returns fixed columns and the shard count.
No checkpoint between batch runs. Write destination: yes (stream; append; option partition_key_column). Streaming source/sink (Enterprise).

MQTT — `mqtt`

Messages published during a collection window.

Key	Required	Default	Notes
`host`	Yes	—
`port`	No	1883 / 8883
`topic`	No	#	Wildcards allowed.
`username / password`	No	—
`tls`	No	false	On whenever a certificate is stored.
`ca_cert_pem`	No	—	CA certificate of a private CA, pasted as PEM text.
`client_cert_pem / client_key_pem`	No	—	Client certificate and key, pasted as PEM text.
`read_seconds`	No	15	Collection window.

Streaming source/sink (Enterprise). Write destination: yes (one topic; append; option qos; up to 200,000 rows per write).

OPC-UA — `opcua`

Current values of variable nodes (one row per node).

Key	Required	Default	Notes
`url`	Yes	—	opc.tcp://host:4840
`node_id`	No	—	Node or folder.
`username / password`	No	anonymous
`client_cert_pem / client_key_pem`	No	—	Client certificate and key, pasted as PEM text.
`security_policy / security_mode`	No	Basic256Sha256 / SignAndEncrypt	Used with the client certificate.
`security_string`	No	None	Policy,Mode,cert,key (self-hosted installs).
`max_nodes`	No	500 / 1000

No history. Write destination: yes (append; node ids and values come from columns; the last row for each node is written; up to 500 rows per write). Needs a saved connection with Allow writing switched on.

InfluxDB — `influxdb`

A measurement over a time range (InfluxDB 2.x, Flux).

Key	Required	Default	Notes
`url`	Yes	—
`token`	Yes	—
`org`	Yes	—
`bucket`	Yes	—
`measurement`	No	—
`range`	No	-30d
flux (API only)	No	—	Complete Flux query.

Write destination: yes (measurement; append, replace; options time_column, required, and tag_columns).

TimescaleDB — `timescaledb`

Tables, hypertables and views (PostgreSQL protocol).

Key	Required	Default	Notes
`connection_string`	Or split fields	—
`host / database / user / password`	Yes*	—
`port`	No	5432
`schema`	No	public
`sslmode`	No	require
`ssl_root_cert_pem`	No	—	Server CA certificate, pasted as PEM text.
`ssl_cert_pem / ssl_key_pem`	No	—	Client certificate and key, pasted as PEM text.

Custom query: yes (API). Incremental: yes. Write destination: yes (replace, append, fail, merge; option time_column for a hypertable).

PI Web API — `pi_web_api`

Recorded values of one PI point.

Key	Required	Default	Notes
`base_url`	Yes	—	https://server/piwebapi
`username / password`	No	—	Basic auth.
`server`	For tag names	first archive
`point`	No	—
`start_time / end_time`	No	-7d /
`verify_ssl`	No	true

Write destination: yes (append, merge; point, time and value columns; up to 500,000 rows per write). Needs a saved connection with Allow writing switched on.

SaaS and apps

Salesforce — `salesforce`

SObject records or SOQL.

Key	Required	Default	Notes
`username / password / security_token`	Password login	—
`domain`	No	login	test for sandboxes.
`consumer_key / consumer_secret`	No	—
`instance_url`	OAuth	—	With consumer key/secret and no username: client-credentials flow.

Custom query: SOQL (API). Write destination: yes (object; append, or merge on one external-id field; up to 150,000 rows per write). Needs a saved connection with Allow writing switched on.

HubSpot — `hubspot`

CRM object records with all properties.

Key	Required	Default	Notes
`access_token`	Yes	—	Private app token (pat-…).
`object`	No	—	Overrides the Discover pick at run time.

Write destination: yes (object type; append, or merge on one unique property; up to 100,000 rows per write). Needs a saved connection with Allow writing switched on.

Stripe — `stripe`

A list resource (charges, customers, invoices…).

Key	Required	Default	Notes
`api_key`	Yes	—	sk_… or restricted rk_…
`object`	No	—	Overrides the Discover pick at run time.

Write destination: yes (creates customers, products and prices; append; up to 10,000 rows per write). Needs a saved connection with Allow writing switched on.

Google Analytics 4 — `ga4`

One report defined by dimensions, metrics and dates.

Key	Required	Default	Notes
`property_id`	Yes	—	Numeric.
`service_account_json`	Yes	—	Viewer on the property.
`metrics`	Yes	—	Comma-separated.
`dimensions`	No	—
`start_date / end_date`	No	30daysAgo / today

Write destination: no.

NetSuite — `netsuite`

SuiteQL tables or queries (token-based auth).

Key	Required	Default	Notes
`account_id`	Yes	—
`consumer_key / consumer_secret`	Yes	—
`token_id / token_secret`	Yes	—
`table`	No	—	Overrides the Discover pick at run time.

Custom query: SuiteQL (API). Write destination: yes (record type; append, or merge on the external id; up to 1,000 rows per write). Needs a saved connection with Allow writing switched on.

SharePoint Lists — `sharepoint`

Items of a list (Microsoft Graph, app-only).

Key	Required	Default	Notes
`tenant_id / client_id / client_secret`	Yes	—	Sites.Read.All.
`site_url`	Yes	—
`list`	No	—	Overrides the Discover pick at run time.

Write destination: yes (list; append; up to 5,000 rows per write). Needs a saved connection with Allow writing switched on.

REST / HTTP API — `http`

One JSON or CSV endpoint, one request per run.

Key	Required	Default	Notes
`url`	Yes	—
`method`	No	GET	GET, POST, HEAD, OPTIONS.
api_key + api_key_header	No	X-API-Key
`api_key_param`	No	—	Send the API key in this query parameter in place of a header.
`bearer_token`	No	—
`token_url / client_id / client_secret / scope`	No	—	OAuth2 client credentials.
`username / password`	No	—	Basic.
`headers`	No	—	JSON.
`body`	No	—	POST only.
`json_path`	No	auto	Dotted path to the rows array.

Private and internal addresses are refused; redirects are not followed. Write destination: yes (append; rows are sent as JSON in batches; options batch_size and method; up to 1,000,000 rows per write).

Connector Reference

Databases

SQL Database — sql

MongoDB — mongodb

Oracle Database — oracle

Apache Cassandra — cassandra

Amazon DynamoDB — dynamodb

Elasticsearch / OpenSearch — elasticsearch

Data warehouses

Snowflake — snowflake

BigQuery — bigquery

Amazon Redshift — redshift

ClickHouse — clickhouse

Databricks — databricks

Microsoft Fabric — fabric

Azure Synapse — synapse

SAP HANA — sap_hana

Storage and files

Amazon S3 — s3

Google Cloud Storage — gcs

Azure Blob Storage — azure_blob

Azure Data Lake Storage Gen2 — adls_gen2

Delta Lake — delta

Google Sheets — google_sheets

Google Drive — google_drive

Dropbox — dropbox

OneDrive / SharePoint Files — onedrive

Box — box