Connectors: what works today, what is next, and what is on the roadmap

Read-only connectors that run inside your agent.

Every connector runs inside the customer-resident agent and reads sources read-only. This list is generated from the capability registry the product uses, and each status is derived from recorded conformance evidence: a connector is only called available when the agent ships it and it passed every check.

Find a connector

Showing all 34 connectors.

Browse by category: Databases · Data warehouses · Lakehouses · Object storage · Files · NoSQL · SaaS applications · Mainframe · Streaming · APIs.

Available now (13)

Shipped in the standard agent and backed by recorded conformance evidence: read-only proof, secret redaction, discovery, bounded streaming, egress canaries and the full discover, mask, subset, certify and provision workflow.

Available now Databases

PostgreSQL

Shipped in the standard agent and backed by recorded conformance evidence.

Built-in connector (psycopg 3): catalog discovery, row-count estimates, column profiles, foreign-key relationships and streaming Arrow reads for subsetting.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: password (secret reference)
Tested environments: PostgreSQL 16.15; PostgreSQL 16.2
Last verified: 2026-09-30 · connector 0.1.0

Details and docs for PostgreSQL · Conformance evidence for PostgreSQL

Available now Databases

Generic SQL (SQLAlchemy)

Shipped in the standard agent and backed by recorded conformance evidence.

Built-in connector for a SQLAlchemy URL: metadata, counts, profiles and streaming reads, with a statement guard that blocks anything other than SELECT/WITH/SHOW/EXPLAIN.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: SQLAlchemy URL (secret reference)
Tested environments: PostgreSQL (SQLAlchemy dialect) 16.2
Last verified: 2026-09-30 · connector 0.1.0

Details and docs for Generic SQL (SQLAlchemy) · Conformance evidence for Generic SQL (SQLAlchemy)

Available now Files

Parquet files

Shipped in the standard agent and backed by recorded conformance evidence.

Built-in file-set connector for Parquet (or CSV) files on a local or mounted filesystem; each declared file or directory is a table.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: filesystem permissions
Tested environments: Local filesystem (Parquet via Apache Arrow) pyarrow 25.0.1
Last verified: 2026-09-30 · connector 0.1.0

Details and docs for Parquet files · Conformance evidence for Parquet files

Available now Files

Local files (Parquet / CSV)

Shipped in the standard agent and backed by recorded conformance evidence.

Built-in file-set connector for Parquet or CSV extracts in a local or mounted directory (for example a read-only landing zone for exports).

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: filesystem permissions
Tested environments: Local filesystem (Parquet via Apache Arrow) pyarrow 25.0.1
Last verified: 2026-09-30 · connector 0.1.0

Details and docs for Local files (Parquet / CSV) · Conformance evidence for Local files (Parquet / CSV)

Available now Object storage

Amazon S3 / S3-compatible storage

Shipped in the standard agent and backed by recorded conformance evidence.

Built-in file-set connector for Parquet or CSV objects under a bucket prefix in Amazon S3 or an S3-compatible store (endpoint override).

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: access key (secret reference); instance/workload identity
Tested environments: SeaweedFS S3 gateway (S3-compatible) 4.48; Amazon S3 service (amazon-s3-service)
Last verified: 2026-09-30 · connector 0.1.0

Known limitations (3)

Verified against Amazon S3 (a real bucket, least-privilege reader) and an S3-compatible server (SeaweedFS S3 gateway); other S3-compatible stores are not individually tested.
Only access-key authentication is verified; instance and workload identity are not yet part of the conformance run.
Read-only access cannot be proven for object stores: it is attested by the operator, and the source is refused without that attestation.

Details and docs for Amazon S3 / S3-compatible storage · Conformance evidence for Amazon S3 / S3-compatible storage

Available now Object storage

Azure Blob Storage / Data Lake Storage

Shipped in the standard agent and backed by recorded conformance evidence.

Built-in file-set connector for Parquet or CSV files in an Azure Blob container or ADLS Gen2 filesystem, with account-key, SAS or service-principal authentication.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: account key; SAS token; service principal
Tested environments: Azure Storage (Blob / Data Lake Storage Gen2) service
Last verified: 2026-10-06 · connector 0.1.0

Known limitations (3)

Verified against a real Azure Storage account with hierarchical namespace (Data Lake Storage Gen2) using a read + list SAS; account-key and service-principal authentication were verified against Azurite, Microsoft's local emulator, only.
SAS tokens as copied from the Azure portal or CLI (without a leading '?') need agent 0.2.0 or later; agent 0.1.0 accepts them only with the leading '?'.
Read-only access cannot be proven for object stores: it is attested by the operator, and the source is refused without that attestation.

Details and docs for Azure Blob Storage / Data Lake Storage · Conformance evidence for Azure Blob Storage / Data Lake Storage

Available now Databases

Microsoft SQL Server

Shipped in the standard agent and backed by recorded conformance evidence.

Dedicated SQL Server connector (pure-Python TDS driver): catalog discovery, primary and foreign keys, row estimates from sys.partitions, aggregate profiles and bounded streaming reads into the local engine, over verified TLS by default.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: SQL authentication (secret reference)
Tested environments: Microsoft SQL Server 2022 (Linux container) 16.0.4295.3
Last verified: 2026-09-30 · connector 0.1.0

Known limitations (4)

Requires agent 0.2.0 or later; agent 0.1.0 does not include its driver.
SQL authentication only; Microsoft Entra ID (service principal) and Windows integrated authentication are not supported yet.
Tested against the SQL Server 16.0 engine (Linux container); Azure SQL Database, Azure SQL Managed Instance and Amazon RDS for SQL Server are compatibility profiles still to be tested.
Geography, geometry, hierarchyid, xml and sql_variant columns are carried as text.

Details and docs for Microsoft SQL Server · Conformance evidence for Microsoft SQL Server

Available now Databases

Oracle Database

Shipped in the standard agent and backed by recorded conformance evidence.

Dedicated Oracle connector (python-oracledb thin mode, no Oracle Client needed): schema discovery, primary and foreign keys, optimizer row estimates, aggregate profiles and bounded streaming reads inside read-only transactions.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: password (secret reference)
Tested environments: Oracle Database Free (container) 23.26.3.0.0
Last verified: 2026-09-30 · connector 0.1.0

Known limitations (4)

Requires agent 0.2.0 or later; agent 0.1.0 does not include its driver.
Password authentication only. The conformance run used plain TCP (the test listener has no TLS endpoint); it proved that the TCPS default refuses such a listener, but a TCPS session and wallet-based mTLS are not yet part of the run.
Tested against Oracle Database Free (23.26); Oracle 19c and Amazon RDS for Oracle are still to be tested.
XMLTYPE and INTERVAL columns are carried as text, DATE as a timestamp, and NUMBER without declared precision as double precision; object, nested-table and VARRAY columns are not read.

Details and docs for Oracle Database · Conformance evidence for Oracle Database

Available now Databases

MySQL

Shipped in the standard agent and backed by recorded conformance evidence.

Dedicated MySQL connector (PyMySQL): information_schema discovery, primary and foreign keys, row estimates, aggregate profiles and unbuffered streaming reads.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: password (secret reference)
Tested environments: MySQL 8.4.11
Last verified: 2026-09-30 · connector 0.1.0

Known limitations (5)

Requires agent 0.2.0 or later; agent 0.1.0 does not include its driver.
Password authentication only (IAM database authentication not yet tested).
Tested against MySQL 8.4 LTS; Amazon RDS for MySQL and Aurora MySQL are compatibility profiles still to be tested.
Spatial values are read and provisioned as WKT text; BIT values as unsigned integers.
The conformance run used TLS without certificate verification against a self-signed test server; verified TLS against a trusted certificate is not yet part of the recorded evidence.

Details and docs for MySQL · Conformance evidence for MySQL

Available now Databases

MariaDB

Shipped in the standard agent and backed by recorded conformance evidence.

MariaDB through the MySQL connector, with its own compatibility record: MariaDB is tested separately and never assumed compatible because a MySQL driver connects.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: password (secret reference)
Tested environments: MariaDB 11.4.13
Last verified: 2026-09-30 · connector 0.1.0

Known limitations (6)

Requires agent 0.2.0 or later; agent 0.1.0 does not include its driver.
Password authentication only.
Tested against MariaDB 11.4 LTS only.
Roles granted to roles (nested roles) are not expanded; such a source is refused.
System-versioned (temporal) tables are read at their current state only.
The conformance run used TLS without certificate verification against a self-signed test server.

Details and docs for MariaDB · Conformance evidence for MariaDB

Available now Data warehouses

Snowflake

Shipped in the standard agent and backed by recorded conformance evidence.

Dedicated Snowflake connector: INFORMATION_SCHEMA discovery, declared keys, row counts from table metadata, bounded aggregate profiles and Arrow result reads processed inside your network.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: key-pair (secret reference); OAuth token (secret reference); password (secret reference)
Tested environments: Snowflake service
Last verified: 2026-10-07 · connector 0.1.0

Known limitations (4)

Key-pair authentication verified against a real Snowflake account; OAuth and password authentication are not part of the recorded run.
Requires agent 0.3.0 or later; earlier agent images do not include its driver.
Snowflake does not enforce declared foreign keys, so relationships may need explicit mappings.
NUMBER(p,0) columns are carried as 64-bit integers; larger values fail the job with a reason code.

Details and docs for Snowflake · Conformance evidence for Snowflake

Available now Lakehouses

Databricks

Shipped in the standard agent and backed by recorded conformance evidence.

Databricks connector: Unity Catalog catalog.schema.table discovery of Delta tables and views through a Databricks SQL warehouse, informational keys and bounded Arrow reads.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: OAuth M2M service principal (secret reference); personal access token (secret reference)
Tested environments: Databricks SQL warehouse (Unity Catalog) service
Last verified: 2026-10-07 · connector 0.1.0

Known limitations (5)

Verified against a real workspace (Unity Catalog, serverless SQL warehouse) with OAuth machine-to-machine; personal access tokens are accepted but not part of the recorded run.
Requires agent 0.3.0 or later; earlier agent images do not include its driver.
Unity Catalog does not enforce foreign keys; informational constraints are read where declared.
Workspace and metastore administrator rights are invisible to the read-only proof: use a dedicated non-admin service principal.
Tags are read locally only; external locations, volumes, Delta Sharing and Jobs are not supported.

Details and docs for Databricks · Conformance evidence for Databricks

Available now Data warehouses

Google BigQuery

Shipped in the standard agent and backed by recorded conformance evidence.

BigQuery connector: dataset and table discovery through the tables API, partition and clustering metadata, row counts from table metadata, and reads through tabledata.list, which is not billed as query scanning.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: service-account key (secret reference); workload identity / application default
Tested environments: Google BigQuery service
Last verified: 2026-10-07 · connector 0.1.0

Known limitations (4)

Verified against a real Google Cloud project with service-account keys. A key is limited to the read-only scope, which cannot run query jobs: tables are read and counted without jobs, but views need application-default credentials (for example the agent's attached service account).
Requires agent 0.3.0 or later; earlier agent images do not include its driver.
BigQuery primary and foreign keys are unenforced; relationships may need explicit mappings.
Profiles are computed by the agent over a bounded sample of listed rows.

Details and docs for Google BigQuery · Conformance evidence for Google BigQuery

Preview (2)

Real, executable support that you can evaluate today, with the limitations listed on each card. A preview connector has passed the minimum conformance checks on a real engine, not all of them.

Preview Mainframe

Mainframe files (EBCDIC / copybook)

Real, executable support with the limitations listed on this page.

Built-in connector for mainframe record files described by COBOL copybooks: EBCDIC code pages, fixed and variable-length records, zoned and packed decimal (COMP-3) and binary fields, decoded inside your network into the same masking and certification workflow.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: filesystem permissions
Tested environments: EBCDIC record files (cp037, FB + VB) with COBOL copybooks IBM record formats F/FB/V/VB
Last verified: 2026-09-30 · connector 0.1.0

Known limitations (5)

Verified with synthetic EBCDIC datasets written by DataNivra's own record encoder; not yet verified against unload files produced on a z/OS system.
Reads unloaded extract files (QSAM/VSAM unloads) from a landing directory; it does not connect to z/OS, VSAM or Db2 for z/OS directly.
REDEFINES, OCCURS DEPENDING ON and floating-point (COMP-1/COMP-2) items are refused with a reason code.
Variable-length (V/VB) files are read in one pass up to a configurable size limit.
Files carry no keys: relationships come from industry-pack templates or explicit mappings.

Details and docs for Mainframe files (EBCDIC / copybook)

Preview Databases

SQLite (developer evaluation)

Real, executable support with the limitations listed on this page.

SQLite database files read through the Generic SQL connector with a verified read-only session, for local developer evaluation of the full workflow.

Verified capabilities: Discovery, Classification, Profiling, Masking, Deterministic masking, Entity-aware subsetting, Relationship preservation, Synthetic workflows, Certification, Provisioning, CI/CD, Evidence generation
Authentication: file permissions
Tested environments: SQLite 3.50.4
Last verified: 2026-09-30 · connector 0.1.0

Known limitations (2)

Intended for local developer evaluation, not for production sources.
A SQLite file has no credentials, so there is no separate read-only login to verify; read-only access comes from the file mode and query_only.

Details and docs for SQLite (developer evaluation)

Coming next (0)

Active implementation. These connectors are being built and tested now, but they are not yet available for production use, and the console does not offer them yet.

No connector currently has this status.

Projected (19)

Strategic roadmap only. Nothing here ships today and no delivery date is committed. Tell us which ones matter to you: demand informs the order.

Projected Streaming