How to choose the right Farcaster data product for your app

Developers can focus on what they are building instead of running a hub and replicator which can be significant cost & effort over time

Product	Pros	Cons
Parquet exports	Developer ingests Apache Parquet files and can read/write to their own db as needed	Need to set up parquet ingestion workflow, open source repo available. See here.
	No need for developer to share access to database w/ Neynar
Hosted SQL	Directly write sql queries against a postgres hosted db, no need to set up anything additional	Developer has no write access to db creating new indexes, etc. requires reaching out to Neynar no changes to overall table schema
Indexer service - pipe Farcaster data	Neynar writes to a db on the developer’s side, no need for developer to manage hubs or replicators Neynar handles all hub and replicator related upgrades and maintenance	Developer needs to have a db that they can share write access to
	Developer has flexibility to let Neynar pipe specific tables instead of all FC data
	Developer can manage the db as they see fit — create new indexes, etc.
Kafka stream	Good real time complement to services like Parquet -- backfill with Parquet and ingest real time with Kafka stream	Need to set up Kafka ingestion. See open source code here.