01 / Integrations

Connect the source, keep the credentials small.

38 native connectors, plus a generic REST, webhook and SFTP path. Every one of them asks for read-only access scoped to a prefix. If a connector needs more than that to work, we treat it as a bug.

02 / Read modes

Four ways in, one rule engine.

The connector decides how rows arrive. What happens to them afterwards does not change: the same compiled rules, the same quarantine store, the same append-only ledger.

CDC

Read from the replication log. Rows are checked before they land downstream.

7 connectors

Stream

Rules run on the consumer side, before the message is committed.

9 connectors

Batch

New objects and partitions are validated on arrival, not on a nightly sweep.

13 connectors

Poll

Fetched on a schedule you set, with provider rate limits respected by default.

9 connectors

Filter

38 connectors

Warehouses

Validation pushed down to the warehouse where the compute already lives.

Snowflake

Batch

Schema drift, null rates, referential gaps
BigQuery

Batch

Partition freshness, type widening, row counts
Redshift

Batch

Sort key skew, null rates, duplicate keys
Databricks

Batch

Delta schema evolution, late arrivals
Synapse

Batch

Distribution skew, type drift
ClickHouse

Batch

Merge lag, cardinality shifts
Firebolt

Batch

Index freshness, null rates

Databases

Change data capture from the replication log. No trigger, no agent in the database.

PostgreSQL

CDC

Constraint violations, orphan rows, enum drift
MySQL

CDC

Charset anomalies, zero dates, orphan rows
SQL Server

CDC

Collation mismatch, null rates
Oracle

CDC

Number precision loss, orphan rows
MongoDB

CDC

Shape divergence across documents, missing fields
DynamoDB

Stream

Attribute type flips, hot partitions
Cassandra

CDC

Tombstone pressure, wide rows
CockroachDB

CDC

Constraint violations, clock skew on ranges

Streams and queues

Rules run on the consumer side, before the message is committed downstream.

Kafka

Stream

Registry compatibility, poison messages, key nulls
Kinesis

Stream

Shard lag, malformed payloads
Pub/Sub

Stream

Redelivery storms, schema mismatch
RabbitMQ

Stream

Dead letter growth, payload shape
Redpanda

Stream

Registry compatibility, poison messages
EventBridge

Stream

Detail-type drift, payload shape

Object storage

Prefix-scoped, read-only. New objects are validated on arrival, not on a nightly sweep.

Amazon S3

Batch

Header drift, encoding, partial uploads
Google Cloud Storage

Batch

Header drift, encoding, object size anomalies
Azure Blob

Batch

Header drift, encoding, partial uploads
MinIO

Batch

Header drift, checksum mismatch
Cloudflare R2

Batch

Header drift, encoding

SaaS and business systems

Polled on a schedule you set, with the provider rate limits respected by default.

Salesforce

Poll

Picklist drift, required-field gaps, duplicate accounts
HubSpot

Poll

Property drift, duplicate contacts
Stripe

Poll

Currency mismatch, orphan subscriptions
Segment

Stream

Tracking plan violations, missing traits
Zendesk

Poll

Custom field drift, orphan tickets
NetSuite

Poll

Subsidiary mismatch, currency rounding
Shopify

Poll

Variant gaps, tax field nulls
Workday

Poll

Effective-dating gaps, PII field tagging
Intercom

Poll

Attribute drift, duplicate users

Generic

For everything that has no connector yet. Same rule engine, same ledger.

REST

Poll

Response shape, pagination gaps, rate limits
Webhook

Stream

Signature verification, replay, payload shape
SFTP

Batch

Header drift, encoding, truncated transfers

03 / Not listed

Your source is not here.

Then use the REST, webhook or SFTP path — same rule engine, same ledger, no waiting list. If the source is common enough that we should have it natively, tell us which one and we will say honestly whether it is on the roadmap or not.

Request a connector Email sales@neuryn.com