> Documentation index: https://docs.beetl.io/llms.txt # Connect your data > The ways data gets into Beetl: integrations, webhooks, files, and data read in place. Source: https://docs.beetl.io/sources/ Last updated: 2026-11-10 Data reaches Beetl in one of five ways. Three of them copy data into a dataset that Beetl writes and keeps up to date. The other two read data where it already sits, so nothing is copied. | Way in | What it is | Data is | Who sets it up | Page | | ------------------ | ------------------------------------------------------------------------------------- | ------------- | -------------------------------------------------------------- | --------------------------------------------------------------------------------- | | Integration | A connection to a system through an installed Connect package, with one or more syncs | Copied | An admin installs the package, then any user connects | [Integrations](https://docs.beetl.io/sources/integrations) | | Webhook | A URL and a token that your system POSTs JSON to | Copied | Any user | [Webhook](https://docs.beetl.io/sources/webhook) | | File Set | A folder of files you upload: CSV, NDJSON or Parquet, or documents and spreadsheets | Copied | Any user | [File Sets](https://docs.beetl.io/sources/file-sets), [Documents and spreadsheets](https://docs.beetl.io/sources/documents) | | Registered dataset | Delta, Parquet, CSV or JSON files already in an object store the tenant uses | Read in place | Any user | [Registered datasets](https://docs.beetl.io/sources/registered-datasets) | | External dataset | A table in a live PostgreSQL database | Read in place | An admin adds the database, then any user registers its tables | [External datasets](https://docs.beetl.io/sources/external-datasets) | All five start on Home. Right-click the canvas and choose **New connection** for an integration, a webhook or a PostgreSQL database, or **Register data set** for files in your object store. File Sets have their own File sets card on the canvas, and **New file processor** starts from a processor that reads the files you add to it. ## Which one to use Use an **integration** when the data lives in a system with an API, such as a SaaS application, and you want Beetl to fetch it on a schedule. Each integration is a Connect package that a TenantAdmin installs in the tenant. If the system you need has no package yet, a developer can write one; see [Build integrations](https://docs.beetl.io/connect). Use a **webhook** when your system can push events as they happen. Nothing needs installing, and nothing of Beetl's runs on your side. Use a **File Set** for exports, one-off loads and anything that arrives as files. CSV, NDJSON and Parquet files are imported as they are. For PDFs, Word files, images and other documents, a file processor reads each document into one record, and for Excel workbooks with a fixed layout you can map cells to columns instead. See [Documents and spreadsheets](https://docs.beetl.io/sources/documents). Use a **registered dataset** when the files are already in your own bucket and you do not want a second copy. Use an **external dataset** when the data sits in a PostgreSQL database and you want every query to see its current state. Each query then runs against your database, so heavy use adds load there. ## How copied data lands A webhook connection, a sync and a File Set's processor each write to exactly one dataset. The dataset shows as Pending on Home until its first successful write, then as Initialized. A run that fails leaves the dataset as it was, and an incremental sync never skips records because of a failed run. New datasets are written to the object store marked as Beetl storage. That is Beetl's hosted store unless an admin has chosen your own bucket instead; see [Tenants and roles](https://docs.beetl.io/concepts/tenants-and-roles#where-your-data-is-stored). The rules behind this, including write modes, partitioning, limits and what deleting a producer does, are on [Connection types, connections and syncs](https://docs.beetl.io/concepts/connections-and-integrations). ## Reaching your systems Integrations and external datasets connect out from Beetl, so the system has to be reachable from the public internet. Integrations reach their providers over HTTPS. A system that only exists inside your private network cannot be reached directly. Export from it into a File Set, or have it push to a webhook. ## In this section | Page | What it covers | | --------------------------------------------------- | --------------------------------------------------------------------------------------- | | [Integrations](https://docs.beetl.io/sources/integrations) | Connections from an installed Connect package: credentials, OAuth, syncs and schedules. | | [Webhook](https://docs.beetl.io/sources/webhook) | Receive JSON events at a per-connection endpoint. | | [File Sets](https://docs.beetl.io/sources/file-sets) | Named folders of uploaded files, turned into datasets by a file processor. | | [Documents and spreadsheets](https://docs.beetl.io/sources/documents) | Turn PDFs, Word files, spreadsheets and other documents into dataset rows. | | [Registered datasets](https://docs.beetl.io/sources/registered-datasets) | Files already in your object store, read in place. | | [External datasets](https://docs.beetl.io/sources/external-datasets) | A live PostgreSQL table, read in place. | ## Related - [Connection types, connections and syncs](https://docs.beetl.io/concepts/connections-and-integrations): The objects behind copied data, and the rules they follow. - [Tenants and roles](https://docs.beetl.io/concepts/tenants-and-roles): What an admin sets up, and where your data is stored. - [Build integrations](https://docs.beetl.io/connect): Writing a Connect package for a system that has none.