Integrations
Plug in anything that emits events.
Connect your queues, databases and SaaS tools in a few clicks. Send typed tables on to your warehouse, dashboards and models without writing glue code.
Catalog
64 integrations
Early access
12 in beta
Typical setup
2-minute setup
Catalog
In, out and everywhere between.
Anthropic
Give Claude live, row-level context through a typed tool.
BigQuery
Write views through the Storage Write API, partitioned by time.
Cloudflare Workers
Query the nearest Basalt region from any Worker.
Grafana
Live panels over the Postgres wire protocol, no plugin needed.
Hex
Query views from SQL cells and keep notebooks current.
Iceberg
Commit views to Apache Iceberg tables with snapshot isolation.
Kafka
Read any topic with consumer-group offsets and exactly-once commits.
OpenAI
Embed events as they land and store vectors beside the rows.
Postgres CDC
Logical replication from the WAL, with snapshots and schema changes.
Segment
Forward track, identify and page calls as typed events.
Snowflake
Sync any view to a table with Snowpipe Streaming.
Stripe webhooks
Signed payment, invoice and subscription events, deduplicated.
Amazon S3
Land Parquet files by hour or by size, with a manifest per batch.
Datadog
Ship lag, throughput and p99 metrics per source and view.
dbt
Run dbt models as incremental Basalt views, tests included.
Jupyter
Load any table into a DataFrame with the Python SDK.
Kinesis
Stream shards from AWS with enhanced fan-out and checkpointing.
LangChain
A retriever and a tool that read Basalt tables directly.
Metabase
Point a Postgres connection at Basalt and build questions live.
MongoDB
Change streams from replica sets and Atlas clusters.
MySQL CDC
Binlog capture for MySQL 8 and Aurora, row-level and ordered.
Pinecone
Upsert embeddings from a view and keep the index in sync.
Redpanda
Kafka-compatible ingest with the same offsets and ordering guarantees.
Vercel
Read tables from Edge Functions with one environment variable.
64 in the catalog · 24 shown here
Connector SDK
Build your own in an afternoon.
The connector SDK is open source under Apache 2.0. Describe the schema, yield events, and Basalt handles checkpoints, retries, back-pressure and exactly-once delivery for you.
sources/inventory.ts
import
_
{ defineSource, t }
_
from
_
"@basalt/connector-sdk"
;
export default
_
defineSource({
__
name:
_
"warehouse-inventory"
,
__
schema: {
____
sku: t.string(),
____
on_hand: t.int(),
____
updated_at: t.timestamp(),
__
},
__
// Basalt stores the cursor and resumes after restarts
__
async *
poll({ cursor }) {
____
const
_
rows =
_
await
_
api.changesSince(cursor ??
_
0
);
____
for
_
(
const
_
row
_
of
_
rows)
_
yield
_
{ key: row.sku, value: row };
__
},
__
every:
_
"5s"
,
});
✓
schema registered · warehouse-inventory v1
✓
first checkpoint committed in 41 ms
✓
source live in eu-north-1
Missing a source?
Tell us what you run. We ship new connectors every two weeks, and requests with the most votes go first.
