Refactor for scalability (0.0.20+) #30
Description
v0.0.19 validated the core concept — decentralized chat on Solid pods works. But to compete with Telegram and WhatsApp, we need to rethink the architecture for serious scale.
Current limitations:
- Chat history — loading full Turtle files doesn't scale to thousands of messages. Telegram handles billions daily. JSS now supports MongoDB, enabling efficient queries over large datasets
- Data layer — JSON-LD native would reduce parsing overhead and open the door to structured queries, search, and bot integrations
- Message storage — sharded architecture for handling large volumes efficiently
- Bundle size — lighter footprint, fewer dependencies, faster load. WhatsApp Web is instant. We should be too
- Interoperability — chat messages as structured linked data means agents, bots, and third-party apps can plug in natively — something neither Telegram nor WhatsApp offer
- Protocol bridges — interop with OpenClaw, Matrix, and Telegram so users aren't locked into a single network
- Privacy first — messages stored on your pod, under your control, on your server. No corporate data mining, no metadata harvesting, no third party reading your conversations
Every existing option has a gap: Signal gives you privacy but you can't export your data. Telegram and WhatsApp give you scale but you can't run the protocol. Matrix gives you federation but it's overcomplicated, slow, and not agent-ready.
Solid chat can be all of these — private, scalable, federated, portable, and agent-native — because the data model is right: structured linked data on pods you own. Plan is to take what we've learned from 0.0.x and refactor the core for that level.