PlayHT 2.0
Audio Retired Full comparison ↗ playht/playht-2 · by PlayHT
· neural-tts
Pricing — 1 offering(s)
About this offering: Retired: PlayHT shut its service down on 2025-12-31, after Meta acquired the company in July 2025 (reported by Bloomberg and TechCrunch), so this model is no longer available. The price shown is its last effective rate on the $49-a-month Creator plan. PlayHT 2.0 was known for expressive speech and voice cloning from a 3-second sample.
Text characters (API)
Retired — no longer sold · last price $30.00 / 1M characters, until 2025-12-31
- $30.00 / 1M characters
Historical
2023-07-01 → 2025-12-31
Effective rate via Creator plan API ($49/month, ~1.6M characters). Closed out: PlayHT was acquired by Meta…
Effective rate via Creator plan API ($49/month, ~1.6M characters). Closed out: PlayHT was acquired by Meta (July 2025) and the service was fully shut down 2025-12-31 — this price is no longer purchasable by anyone. See the playht provider record for shutdown sourcing.
Showing the active price and any recorded history. Full pricing history is available via the paid API — see API docs.
See how PlayHT 2.0 fits into a cost-aware routing setup
See how →Capability profile
How PlayHT 2.0 rates across core capability dimensions, with the task-level evidence behind each rating.
Consistently rated among the most natural TTS outputs in community comparisons; strong expressiveness and prosody control.
142 languages supported — one of the widest coverages among dedicated TTS providers.
800+ preset voices plus instant and professional voice cloning.
Streaming (WebSocket) supported. PlayHT 2.0 Turbo variant offers lower latency for real-time applications.
Instant Voice Cloning from 3 seconds of audio; Professional Clone from 1–30 minutes for higher fidelity brand voices.
Benchmarks
| Benchmark | Score | Config | Source |
|---|---|---|---|
| MOS (naturalness) | — | No independently published quality score; PlayHT doesn't appear in the Artificial Analysis Text to Speech Arena (checked 2026-08-19). | — |
Operator guidance
The best option when voice cloning from short samples matters and the ElevenLabs credit-based pricing is limiting. At ~$30/1M chars (effective Creator plan rate) it sits between OpenAI TTS-1 HD and ElevenLabs Flash v2.5 on price. 142 languages is broader than ElevenLabs (29–32). For ultra-low-latency real-time applications, Cartesia Sonic's SSM architecture is faster. Verify current PAYG pricing at the pricing page before committing.
Use cases
- Voice cloning for consistent brand voices from short reference audio
- High-quality multilingual TTS across 142 languages
- Emotionally expressive voiceover and narration content
- Applications requiring instant custom voice creation without enterprise contracts
Limitations
- PAYG API pricing is not prominently published; verify current rate before use
- Credit-based plan pricing; effective per-character cost is plan-dependent
- Professional Clone voice quality requires 1–30 minutes of reference audio