Chat TTS

AI voice synthesis for your chat

Turn viewer messages and channel point redemptions into natural AI voices. Real-time queue, multiple engines, and direct OBS delivery.

<200mssynthesis latency

3synthesis engines

∞voice assignments

Features

Channel Points trigger

Bind TTS to a specific channel points reward. Only matching redemptions are synthesized—the bot ignores everything else automatically.

Choose from Qwen3-TTS, OmniVoice, and Edge-TTS backends. Each offers different quality profiles and latency characteristics.

Messages queue up and play sequentially. Skip items, clear the queue, or pause TTS from the moderator panel—live, mid-stream.

Assign a unique voice to frequent viewers or subscribers. Their messages always play in their designated voice.

Synthesized audio goes straight to your OBS browser source over WebSocket. No disk writes, no polling, near-zero latency.

URLs, mentions, and special characters are stripped automatically. TTS stays readable and natural without manual filtering.

Multiple voice personas available. Each can be assigned per-user or set as the channel default.

Natural

Warm conversational tone, great for casual chat

Calm

Clear and measured, ideal for announcements

Energetic

Lively and bright, matches high-energy streams

Soft

Gentle and smooth, low-fatigue for long streams

01Trigger

A viewer uses the configured channel points reward and types their message. The SidePilot bot captures the redemption immediately.

02Synthesis

The backend picks the correct voice, sends the text to the synthesis engine, and streams the audio back within milliseconds.

03Broadcast

The audio is delivered to your OBS browser source in real time. Viewers hear it the moment synthesis completes.

Turn viewer messages and channel point redemptions into natural AI voices. Real-time queue, multiple engines, and direct OBS delivery.