VoiceMimicry

Status is working but tested in singleplayer This mod is currently undergoing active validation. Technical instabilities may occur(or not?).

Core Functionality Mod(Plugin) makes monsters copy voice and play it during game with memory function what resets with each level.

Technical Specifications Raw Stream Hooking: Hooks directly into the PushDataAsync Photon Voice pipeline. This ensures high-fidelity audio capture without the "slow-motion" or "distortion" bugs found in traditional Unity microphone implementations. Zero Disk I/O: Audio data is stored exclusively in RAM(don't mix up with disk) as compressed 16-bit PCM arrays. It never writes to the SSD, preventing stuttering and hardware wear. Staggered Networking: Uses a custom coroutine-based chunking system. Large voice clips are split and sent across multiple frames to stay under the Photon 100k element limit and prevent TimeoutDisconnects. Memory Management: The voice bank is automatically purged upon every level transition (LevelGenerator.Generate) to prevent memory leaks and cross-level audio contamination.

Mimicry Logic Clip Aging: Clips are locked for 10 seconds immediately after recording. Entities cannot play a clip until it has matured, preventing "instant echo" mod detection. Proximity Prioritization: Entities calculate the distance to all active players and prioritize playing the voice of the closest target to maximize psychological impact. Dynamic Configuration: Monster names are parsed from the EnemyParent root. Config entries are generated only for valid networked entities. Generic folders ("Enable", "Controller") are hard-filtered.

Config("BepInEx/Config/com.randomlygenerated.voicemimicry.cfg") Adjust MinDelay, MaxDelay, and Volume via the BepInEx configuration file. Individual enemies can be toggled on/off dynamically.

Even if I spend day on writing this mod(+ testing) I think it's works well in singleplayer. If something don't works just go in manager and delete this mod(or use my patcher, patcher works better I guess).