Integrations
Arc works with the tools you already run
Arc takes data in over the protocols your agents already speak, writes it as open Parquet on storage you own, and publishes it as Iceberg tables. Here is each integration, how it connects, and where to read more.
Collect
Get data in with the agents you already run
Arc speaks InfluxDB Line Protocol and its own columnar MessagePack format, and ships native plugins for the collectors and stream processors that already move your telemetry.

Telegraf
Telegraf's native Arc output plugin sends columnar MessagePack, gzip-compressed, so any of its 300+ inputs, OPC UA and Modbus included, can write to Arc.
Line Protocol
Arc accepts InfluxDB Line Protocol on InfluxDB's own write paths, so InfluxDB client libraries and agents point at Arc without code changes. Bulk .lp files import too.

MessagePack
Arc's recommended write format: typed, columnar batches posted to /api/v1/write/msgpack.
MQTT
Arc subscribes to your brokers itself, with no bridge in between. Subscriptions are managed over the REST API or in Launchpad, with JSON or MessagePack payloads.
Node-RED
The @basekick-labs/node-red-arc nodes write batches to Arc and run SQL queries from a flow. Node-RED 4.0+.

Redpanda Connect
A native output plugin that speaks Arc's MessagePack protocol, zstd-compressed, from any Redpanda Connect pipeline.
TLE
Arc parses satellite two-line element sets natively: stream them to /api/v1/write/tle or bulk-import a whole catalog.
Store
Open files on storage you own
Arc writes Apache Parquet to local disk or to object storage in your account, and describes those files as Apache Iceberg tables. Nothing about the data is proprietary.
Apache Parquet
Every file Arc writes is Apache Parquet, on every storage backend, readable by any Parquet tool. Existing Parquet files import with one API call.
Apache Iceberg
Arc publishes its own Parquet files as Iceberg tables without copying them (26.09.1+). From 27.01.1 it also serves an Iceberg REST catalog and registers the tables in AWS Glue.
Amazon S3
Store Arc's files in an S3 bucket in your account, with access keys or an IAM role. Any S3-compatible object store works too.
Google Cloud Storage
Arc 27.01.1Store Arc's files in a GCS bucket, with Application Default Credentials or a service-account key file.
Azure Blob
Store Arc's files in Azure Blob Storage, with an account key or a managed identity.

SeaweedFS
An open-source, S3-compatible object store for running Arc on your own servers. Arc's Helm chart ships with it.
Query
Query from code, or straight off the files
Ask Arc with SQL from Python and dataframes, or skip Arc entirely: open engines read its Parquet and Iceberg tables directly.
DuckDB
Arc's SQL is DuckDB's SQL. Outside Arc, DuckDB reads Arc's Parquet files directly, or its Iceberg tables with iceberg_scan.
Polars
The Python SDK's query_polars() runs SQL on Arc and returns a Polars DataFrame.
pandas
The Python SDK's query_pandas() runs SQL on Arc and returns a pandas DataFrame.
Python SDK
arc-tsdb-client covers writes, SQL queries, retention and admin from Python 3.9+.
Arrow Flight SQL
Arc 27.01.1An optional, read-only gRPC listener that returns query results as Arrow record batches to Flight SQL and ADBC clients.
Apache Spark
Spark reads Arc's Iceberg tables in place, by path or, from 27.01.1, through Arc's REST catalog with Iceberg's SparkCatalog.
Trino
Arc 27.01.1Trino's Iceberg connector loads tables from Arc's REST catalog with a read-only token and reads the Parquet from your object store. Runs fully on-prem.
Read in place
Warehouses that read Arc without a copy
From Arc 27.01.1, Arc serves its own Iceberg REST catalog and registers its tables in AWS Glue. Warehouses read the Parquet where it lies: no export job, no load, no connector from us.
Databricks
Arc 27.01.1Unity Catalog reads Arc's tables through an AWS Glue foreign catalog. Arc keeps the Glue tables current.
Snowflake
Arc 27.01.1A catalog-linked database on Arc's own REST catalog, or on Glue, reading Arc's files through a read-only external volume.
Amazon Athena
Arc 27.01.1Arc's measurements appear as tables in Athena's AwsDataCatalog. No crawler, no CREATE TABLE.
Amazon Redshift
Arc 27.01.1Redshift Serverless mounts the Glue Data Catalog on its own, so Arc's tables are queryable with three-part names.
Google BigQuery
Arc 27.01.1A Glue federated dataset. BigQuery Omni runs the query in AWS, next to Arc's files, and only the results cross clouds.
Visualize
Dashboards, BI and the editor
Point the tools your team already opens every day at Arc.
Grafana
The Arc data source plugin runs SQL with Grafana macros, over Arrow, MessagePack or JSON. Grafana 12.3+.
Apache Superset
arc-superset-dialect is a SQLAlchemy driver: add Arc as a database with an arc:// URI.
VS Code
The Arc Database Manager extension: connections, SQL IntelliSense, result grids and CSV import, from the editor.
Power BI
A OneLake shortcut over Arc's Iceberg tables lets Fabric and Power BI read Arc's Parquet in place (27.01.1). A Fabric notebook can also pull query results over Arrow.
Not on the list?
Anything that reads Parquet or Iceberg, or speaks HTTP, can work with Arc: the storage format and the API are both open. If you need something specific, open an issue or ask on Discord.
Ready to handle billion-record workloads?
Deploy Arc in minutes. Own your data in open files on your storage. Use for analytics, observability, AI, IoT, or data warehousing.