Usually, to set up a Firehose environment you need:
- A Firecore binary, which spins up the different components needed (reader, merger, relayer...).
- A Firehose chain-specific binary, which is used as a bridge between Firecore and the blockchain.
- A full instrumented node or an RPC to fetch the blockchain data.
Firecore allows you to spin up all the different Firehose component needed, such as the reader or the relayer. The Firecore is a binary exported in the form of a CLI, so you can easily interact with it.
You can download the latest version of Firecore from the release page of GitHub.
Every chain has a different data model (e.g. Ethereum vs Solana), therefore, Firehose needs to extract the data of every blockchain differently. In order for Firehose to support a new chain, a new extraction layer must be created. This extraction layer is then used by Firecore to interact with the different base components.
Essentially, Firecore acts as the coordinator and uses the chain-specific binary. Some examples of chain-specific binaries are firehose-ethereum or firehose-solana.
Firehose is only an extraction engine, so you still have to provide a valid source of data. There are two modes in which Firehose can extract data:
- Instrumented Node: in this case, you run a full blockchain node (e.g. geth for Ethereum). Note that this requires the node to share the same instrumentation interface used by Firehose, so that it is possible to extract all the data required.
- RPC Poller: in this case, you are DO NOT run the full node, but rely on the RPC API of the blockchain, which usually offers you a poorer data model.