farseer: Another kind of Farcaster hub

Submit
github.com

farseer: Another kind of Farcaster hub

AI summary of the linked article

farseer is a lightweight re-implementation of a Farcaster hub that does not require syncing to operate and can be extended through plugins. Users can bring their own database, logic and infrastructure to harvest protocol data. The project offers a Docker setup that starts the hub with default behavior and plugins, or a manual route that requires Go 1.22+ to compile the binary and plugins. Configuration is handled through a config.toml file, which sets the public IP, gossip port, bootstrap peers, and per-plugin options such as the PostgreSQL plugin's database address, allowed message types and tracked FIDs. Plugins export a Handler struct with functions for message types such as cast adds and reaction adds.

avatar
@noctis, please tell us what your goals are with Farseer. Where do you see this project in 1-5 years?
In 1-5y, I would love the project to have community-made plug-ins who bring a lot of new possibilities to the FC ecosystem. More support for DBs, protocol, etc... I have a great basis but without community support it's hard to support every DB.

My goal is really to create like an "awesome-list" for my project, and try to expand the capabilities of farseer through open-source.

Right now, it's so easy to deploy so anyone could have a cheap powerful hub with FC data available. If I succeed, this could be so beneficial for the network & the developers alike.
avatar
Where do you see the benefit over your solution vs the replicator on Warpcast? I guess your‘s is going to be more neutral in terms of the other tools you can connect to it?
I guess it’s always better to be in control of your data source if you truly want to follow the philosophy of decentralization.

I think that Farseer is not particularly attractive in an ecosystem that provides a lot of Infrastructure-As-A-Service stuff (i.e Neynar) but I definitely think there’s a place for Farseer for those who want to own their data.
avatar
What‘s your definition of owning your data? To me it seems like you equate it to the right to back your data up, for example. Any other rights you think belong in the bucket we call „owning your data?“
For me, owning your data is self-hosting: being able to access it without a third-party involved.