Bark
Suno's expressive transformer-based TTS.
Open Source 8–12 GB VRAM
- Min VRAM
- 8 GB
- GPU class
- Entry GPU
- Quant
- FP16
Actually FreeNo SignupOpen SourceWatermark-Free
XTTS v2 (Coqui)
Multilingual voice cloning in 6 seconds.
Open Source 4–6 GB VRAMRuns locally
Actually FreeNo SignupOpen SourceWatermark-FreeHobbyist-FriendlyAPI
Visit XTTS v2 (Coqui)Updated 2025-09-30 · Direct link
Hardware requirements
Runs locally · Entry GPU (6–8 GB)
Min VRAM
4 GB
Rec. VRAM
6 GB
Min RAM
8 GB
Rec. RAM
16 GB
Disk
4 GB
GPU class
Entry GPU
11.7+Apple Silicon ✓CPU-CapableQuant: FP16
Real-time on 4 GB+. Apple Silicon MPS works.
Screenshot placeholder · XTTS v2 (Coqui)
What is XTTS v2 (Coqui)?
Coqui's XTTS v2 is the production TTS workhorse: clone a voice from 6 seconds of audio, generate speech in 17 languages, run on a 4 GB GPU. Coqui the company is gone but the model lives on under a permissive licence, and it's the backbone of most current OSS voice apps.
Pros & cons
Pros
- 6-second voice cloning that actually works
- 17 languages including cross-lingual cloning
- Real-time on a single mid-range GPU
- Used as the engine inside many higher-level apps
Cons
- Coqui (the company) shut down — community-maintained from here
- Licence is permissive but not OSI-approved; check before commercial use
What's actually free?
Coqui Public Model Licence — free for personal & commercial use.
✓ Actually FreeNo SignupOpen SourceWatermark-Free
Alternatives
Tortoise TTS
Slow, but the quality is worth the wait.
Open Source 4–8 GB VRAM
- Min VRAM
- 4 GB
- GPU class
- Entry GPU
- Quant
- FP16
Actually FreeNo SignupOpen SourceWatermark-Free
F5-TTS
Zero-shot voice cloning TTS — 15 s of audio is enough.
Open Source 8–12 GB VRAM
- Min VRAM
- 8 GB
- GPU class
- Mid GPU
- Quant
- FP16
Actually FreeNo SignupOpen SourceWatermark-Free