Your client.
Your choice of node.
Language-agnostic engineering documentation for a local-first Farcaster client architecture: canonical protocol data, replaceable read providers, explicit capability routing, offline persistence and independently signed writes.
Provider-neutral architectureAny language · Any interfaceGetting started
Understand the product and learn to connect your preferred read node.
02Architecture and flowcharts
Explore domain models, local storage, asynchronous refresh and signed write routing.
03Provider switching
Configure raw Snapchain nodes and indexed APIs with explicit fallback consent.
04Client portability
Understand FIDs, account keys and how different Farcaster clients share protocol records.
05Self-hosting a node
Explore hardware requirements, synchronization, endpoint health and secure operation.
06Implement the architecture
Implement capability ports, provider adapters, persistence boundaries, sync and a durable signing pipeline in your own stack.
A client should not be a single-API wrapper.
Genycaster defines capabilities independently from vendors. Raw protocol reads, indexed social data and write-capable validators are separate routes. UI surfaces consume normalized local state and request bounded background refreshes.
Documentation, not an application mirror
This site documents architecture and backend contracts, not a specific UI toolkit. It is not an application mirror or a hosted API. Capability availability and integration limits are covered in the feature status.