Paxa Labs and Resemble AI for English, Thai, and Mandarin speech
A hosted speech API for English, Thai, and Mandarin next to Resemble AI's open Chatterbox models, across availability, hosting, languages, Thai and Mandarin speech, delivery tags, and voice cloning.
Resemble AI no longer sells generative voice, and it takes no new voice customers. Chatterbox and its other voice models stay open and free to build on, and the Chatterbox repository is now an MIT-licensed archive that is no longer maintained. Building on Resemble AI's voice models now means running Chatterbox on your own hardware. Paxa TTS Flash v2 speaks English, Thai, and Mandarin, and all 50 of its voices speak all three.
Sources:Resemble AI voice archive(opens in a new tab)Resemble AI on deepfake detection(opens in a new tab)Chatterbox repository(opens in a new tab)
Rates and claims last checked October 2026
What differs
| Dimension | Paxa Labs | Resemble AI(opens in a new tab) |
|---|---|---|
| Buying it | A hosted API. Paxa TTS Flash v2 bills per character at one published rate for English, Thai, and Mandarin. | Generative voice is no longer sold, and new voice customers are not taken. Their archive keeps the hosted platform's API guides for reference only. |
| Where it runs | On our hosted API, over chunked HTTP streaming or a live WebSocket. | On your own hardware. Chatterbox installs as a Python package, and its repository is retired research, no longer maintained. |
| Language coverage | English, Thai, and Mandarin, in all 50 voices on one model. | Chatterbox-Multilingual lists 23 languages, Chinese among them. Turbo and Nano speak English only. |
| Thai speech | All 50 voices speak Thai, each with a published id and a fixed sample line. Paxa TTS Flash, model paxa-tts-flash-v1, carries 4 more in 3 regional registers. | Thai is absent from the 23 languages Chatterbox-Multilingual lists. |
| Mandarin speech | All 50 voices speak Mandarin from Simplified Chinese text, in the same voice they use for English and Thai. | Chinese (zh) on Chatterbox-Multilingual, and a dedicated Chinese finetune in its single-language pack. |
| Delivery tags | On Paxa TTS Flash v2, tags written into the text, such as [whispering] or [laughs], direct the delivery. | Turbo reads paralinguistic tags such as [cough], [laugh], and [chuckle], in English. |
| Custom voices | No cloning. A voice can be designed from a written description in the console and called by its id on Paxa TTS Flash v2. | Zero-shot cloning from a reference clip. Every generated file carries a Perth watermark. |
Sources:Resemble AI voice archive(opens in a new tab)Resemble AI on deepfake detection(opens in a new tab)Chatterbox repository(opens in a new tab)
- Paxa TTS Flash v2, per 1M characters
- $10
- Resemble AI
- No published rate
Resemble AI publishes no voice rate. Generative voice is no longer sold, and Chatterbox runs on hardware you provide. Every rate is per one million characters of English text, counted in UTF-16 code units as ours is, with per-minute, per-token, per-byte, and per-Han-character rates converted to that unit.
When Resemble AI is the better choice
If the model has to run on your own machines, Chatterbox is MIT licensed and installs as a Python package. It clones a voice from a reference clip, which we do not offer, and watermarks every file it makes. Existing Resemble AI voice customers are still supported. Choose us when you want a hosted API with Thai voices and a published rate.
Questions
Is Resemble AI still selling text to speech?
No. Their research archive says Chatterbox and their generative voice models remain open and free to build on and are no longer offered as paid products. Their post on moving to deepfake detection says they support existing voice customers and take no new ones.
Does Chatterbox speak Thai?
Thai is absent from the 23 languages its Multilingual model lists. Chinese is on the list, and its single-language pack adds a Chinese finetune.
Do you offer voice cloning?
No. Catalog voices are published with stable ids, and a voice of your own is one you design from a written description in the console. Chatterbox clones from a reference clip when you run it yourself.
Do you publish quality or latency comparisons?
No. We publish no quality, accuracy, or latency figures of our own, for our model or anyone else's. Generate the same script on both and listen.
Other comparisons
- Best Text to Speech APIs for Each Job
- Paxa Labs and ElevenLabs for English, Thai, and Mandarin speech$40-80
- Paxa Labs and Botnoi Voice for English, Thai, and Mandarin speech$315-609
- Paxa Labs and iApp for speech and Thai documents$68-95
- Paxa Labs and Google Gemini TTS for English, Thai, and Mandarin$30-34
- Paxa Labs and OpenAI text to speech for English, Thai, and Mandarin$15-30
- Paxa Labs and Azure AI Speech for English, Thai, and Mandarin text to speech$15-22
- Paxa Labs and Amazon Polly for English, Thai, and Mandarin text to speech$4-100
- Paxa Labs and Deepgram Aura for English, Thai, and Mandarin text to speech$15-30
- Paxa Labs and Murf AI for English, Thai, and Mandarin speech$10-30
- Paxa Labs and the Speechify API for English, Thai, and Mandarin speech$6-10
- Paxa Labs and Narakeet for English, Thai, and Mandarin speech$56-223
- Paxa Labs and Cartesia for English, Thai, and Mandarin speech$50
- Paxa Labs and MiniMax for English, Thai, and Mandarin speech$60-100
- Paxa Labs and Fish Audio for English, Thai, and Mandarin speech$15-50
- Paxa Labs and UVoice AI for English, Thai, and Mandarin speech$111-199
- Paxa Labs and WellSaid for English, Thai, and Mandarin speechNo published rate
Start with free credits
Sign in with Google, GitHub, or Hugging Face and spend 100 one-time free credits on your own text.
Start free