Migrating from LavaFlow

Raya started from LavaFlow’s code and was rebuilt for correctness and resilience. Most concepts map directly.

API changes

LavaFlow	Raya
`new Manager({ nodes, send })` + `manager.init(id)`	`new Raya({ nodes, connector })` (or `send` + `init`)
`defaultSearchPlatform: 'spsearch'`	`defaultSearchSource: 'spsearch'`
`autoPlay: true`	`playerDefaults: { autoplay: true }`
`debug: true`	`debug: true` (same)
`manager.updateVoiceState(packet)`	`raya.handleRaw(packet)` (the old name still works)
`manager.create(options)`	`raya.createPlayer(options)` or `await raya.join(options)`
`manager.get(guildId)`	`raya.getPlayer(guildId)`
`manager.search(query)`	`raya.search(query)` or `player.search(query)`
`player.addTrack(t)` / `addTracks(ts)`	`player.queue.add(t \| ts)` or `player.enqueue(result)`
`player.setVolume(0-100)`	`player.setVolume(0-1000)`, where 100 is normal
`player.setLoopMode(mode)`	`player.setLoop(mode)`
`player.filters().bassboost().apply()`	`await player.filters.bassBoost()`
`player.moveTrack` / `swapTracks` / `removeTrack`	`player.queue.move` / `swap` / `remove`
`player.saveQueue()` / `restoreQueue()`	`raya.shutdown()` / `new Raya({ restore })`
`result.loadType` + `result.data`	`result.type` + `result.tracks`
`manager.destroyPlayer(guildId)`	`player.destroy()` or `raya.destroyPlayer(guildId)`

LavaFlow	Raya
`setVolume(100)` sent 1000, ten times too loud	100 means unchanged
Idle nodes were disconnected after about 60 s	WebSocket pings; idle nodes stay connected
Resuming sent the v3 `Resume-Key` header, so it never worked	v4 `Session-Id` resuming
A global `NodeManager` meant separate managers shared nodes	Each `Raya` owns its nodes
`play()` could run before the node had a session	Players only use ready nodes
The health check permanently removed reconnecting nodes	Nodes reconnect with backoff, and players fail over
Used `fetch` while claiming Node 16 support	Requires Node 18.17+