SoulX-Podcast¶
Tier 6: 1,956 points ยท per 4,000 characters
Comparison method and assumptions.
Pricing¶
Website Pricing¶
| Type | Cost |
|---|---|
| Base | 489 ($0.015) / 1k chars |
| Dialect | 489 ($0.015) / 1k chars |
All parsed rates (both currencies):
| Service | Currency | Amount | Per |
|---|---|---|---|
| Base | USD | 0.015 | 1000 characters |
| Base | POINTS | 489 | 1000 characters |
| Dialect | USD | 0.015 | 1000 characters |
| Dialect | POINTS | 489 | 1000 characters |
Last Checked: 2026-10-22 13:13:48.978692
Bot Information¶
Creator: @empiriolabsai
Description: SoulX-Podcast is an open-source AI voice model developed by Soul AI Lab designed to generate realistic, long-form, multi-speaker podcast dialogues. It features advanced paralinguistic controls that simulate natural human elements like laughter and sighs, while supporting zero-shot voice cloning across Mandarin, English, and various Chinese dialects.
Learn more: https://soul-ailab.github.io/soulx-podcast/
This model is hosted by EmpirioLabs.ai and is exclusive to the Poe platform.
Notes: - Generations may take up to 30 minutes to complete. - Generations are capped at 20k words (~2 hours of audio). - Large audio files (>50mb) will be linked instead of previewed in the Poe message field due to their size.
Parameter controls available:
1. Voice Configuration
- --voice_s\[1-4\] \[arthur/james/lj/xiaomei/zhigang\]: Selects preset voice for specific speaker number
- If cloning voice: Select Custom Voice option for speaker and upload audio
- Model Settings
--model \[base/dialect\]: partial to 'base' (English/Mandarin) or 'dialect' (+ Sichuan, Henan, Cantonese)-
--output_format \[mp3/wav\]: 'wav' for lossless quality, 'mp3' for smaller file size -
Advanced Generation
--temperature \[0.1-2.0\]: Controls randomness in generation (default 0.6)--repetition_penalty \[1.0-2.0\]: Penalizes repetitive output (default 1.25)-
--seed \[number\]: Number for reproducible results (default 42) -
Sampling Controls
--top_k \[1-500\]: Limits vocabulary to top K probable tokens (default 100)--top_p \[0.1-1.0\]: Limits vocabulary to cumulative probability P (default 0.9)
Extra: Powered by a server managed by @empiriolabsai. Learn more
Architecture¶
Input Modalities: Unknown
Output Modalities: Unknown
Modality: unknown
Technical Details¶
Model ID: SoulX-Podcast
Object Type: bot
Created: Unknown
Owned By: empiriolabsai
Root: SoulX-Podcast
API Last Updated: Not listed in API