> ## Documentation Index
> Fetch the complete documentation index at: https://docs.moralis.com/llms.txt
> Use this file to discover all available pages before exploring further.

# What Are Data Feeds?

> Moralis Data Feeds give you the complete, decoded onchain dataset as a real-time stream you own, consumed through whatever already fits your stack.

Moralis **Data Feeds** give you the complete, **decoded onchain dataset**, every block turned into balances,
transfers, swaps, prices, approvals, NFT activity, and raw logs, as a **real-time stream you can own**. It's
the same data that powers Moralis, available for you to consume directly.

## One data lake, many ways out

Data Feeds is a single service over one data lake. Moralis writes each chain's data **once**, then exposes that
same data through whatever interface fits your stack, so you adopt it **without re-tooling**:

### Stream it with the tools you already use

Point your **existing clients** straight at Data Feeds, they're wire-compatible and work unmodified:

* **Kafka**, your Kafka producers/consumers connect as-is.
* **AMQP / RabbitMQ**, standard AMQP 0-9-1 clients.
* **Amazon SQS**, SQS-compatible HTTP API.

…or read directly over **REST / HTTP** or **Arrow Flight (gRPC)** when you want high-throughput columnar pulls.

### Land it with the sink

Point a **recipe** at your own **Postgres, MySQL, or ClickHouse** and the sink continuously lands
ready-to-query tables. Moralis does the heavy lifting in the recipe, decoding, enrichment, and reorg
correction, so you get a production-grade dataset you own and query however you like. This is the smoothest
path for request/response needs (e.g. "a wallet's balances").

### Query it from your warehouse

Every feed is an Apache **Iceberg** table, committed as data lands, so your warehouse and notebooks read it
**directly, without moving data**: **Snowflake, BigQuery, Spark, Trino, Databricks, DuckDB, Pandas, dbt**.

> It's the **same data lake** under all of these, they're access modes of one service, not separate products.
> Choose by your stack and use case; you're never locked into one shape.

## What's in a feed

Each record is a **block**, carrying the decoded onchain activity for that block as typed arrays, token
transfers, swaps, price updates, balances, NFT transfers, approvals, and the raw block/transaction/log data.
See [The `block` array](./the-block-array) for how transactions and logs are laid out.

A few properties worth knowing:

* **Decoded, not raw RPC.** Swaps, transfers, balances, and prices are already extracted, you don't decode
  logs yourself (though the raw logs are there if you want them).
* **Real-time and historical.** Read the latest blocks as they land, backfill history, or both.
* **Reorg-aware.** Each block records its position and parent, so consumers (and the sink's recipes) correct
  for chain reorganizations rather than serving stale data.
* **Onchain data.** A feed contains what's on the chain. Off-chain signals (token logos, spam/verification
  labels, security scores) are layered on separately, see
  [Data Feeds vs. the legacy API](./data-feeds-vs-the-legacy-api).

## Where to go next

* **Migrating from a Moralis REST endpoint?** Start with the relevant migration guide.
* **Building something new?** Browse the recipe cookbook for a ready-to-run feed.
* **Choosing how to consume it?** Skim "One data lake, many ways out" above and pick the interface you already
  use.