TTS Audio Suite, the multi-engine text-to-speech node pack for ComfyUI, shipped version 5.0.0 on June 14, 2026. The release adds Higgs Audio v3, moves the main environment to Transformers 5, and introduces Runtime Isolation so older, dependency-fragile engines stop breaking the rest of your install.

How to Integrate It Into Your ComfyUI Pipeline

Update the node pack through ComfyUI Manager or a git pull, then drop the new Higgs Audio v3 engine into a TTS Text or TTS SRT node. Higgs v3 gives you zero-shot voice cloning from a short reference clip plus inline emotion, style, prosody, and sound-effect tags written directly in the script, so you can direct a performance without leaving the prompt. Because the suite keeps SRT timing and character support across all engines, you can clone a narrator voice, tag it for tone per line, and render a fully timed voiceover track in one graph. Runtime Isolation means you can keep using a legacy engine like an older Chatterbox build in a secondary runtime while the main suite runs on the modern stack.

Why It Matters for Creators

Local TTS stacks rot fast: one engine pins an old Transformers version and the whole node pack stops importing. By splitting modern and legacy engines into separate runtimes, v5 ends the all-or-nothing dependency trap that has plagued ComfyUI audio setups. For creators who want voice work that stays on their own machine, with no per-character API bill, a stable 11-engine suite that now includes a flagship cloning model is a meaningful upgrade.

Key Details

Version: 5.0.0, released June 14, 2026.

New engine: Higgs Audio v3 with zero-shot cloning and inline tags.

Architecture: Transformers 5 main environment plus Runtime Isolation for legacy engines.

Engines: Eleven total, including F5-TTS, VibeVoice, IndexTTS-2, CosyVoice3, Qwen3-TTS, and RVC voice conversion.

What to Do Next

Back up your current node folder before updating, since the architecture shift is large, then rebuild one simple TTS graph with Higgs v3 to confirm your environment resolves before migrating production workflows.