The verdict
SoulGen has repositioned significantly: the live site title is "SoulGen: Free AI Video Generator — Create Talking Videos from Images & Text," and the video engine leads. The model architecture has four named tiers confirmed from the live pricing page: SoulGen 2.0, SoulGen 3.0 (current main), SoulOmni 2.0, and SoulOmni 2.0 Turbo, each with distinct credit costs — SoulGen 3.0 charges 15 credits per 5-second clip (audio off) or 17 with audio on; SoulOmni 2.0 Turbo tops out at 40 credits per 5-second 1080P clip. Reference to Video feeds a reference clip and prompt together to control motion direction. Video Extend and Video Edit each have their own credit tables. DreamTwin Character is a distinct creation mode that generates a styled AI persona from your reference photos — separate from the Face Swap tool, which maps a face consistently across standalone generations. Human Modeling produces photorealistic generated models at 3 credits per image. On the image side: Realistic and Anime Character styles, Classic Edit (inpainting), Extend Image (outpainting), and VR Face Swap for spatial media. Soul Chat integrates companion chat, image generation, and voice calls in the same session (0.2 credits per text message, 2 per image, 1 per minute of voice). A PWA is available — install from the browser on any device without an App Store. The credit economy is the key constraint: at 15 credits per 5 seconds, the 100-credit Pro monthly allotment buys roughly 33 seconds of SoulGen 3.0 video. For higher image volume per dollar, Mage.space or Promptchan offer more generations; for companion depth, SoulGen skews image-heavy.
Best for NSFW creators who want images, video, and a companion in one tool. For higher image volume per dollar, Mage.space (more credits and model control). For images embedded in a deep chat companion, Candy AI.
Pros
- Realistic and anime images, both done well
- Four named video model tiers (SoulGen 3.0 / SoulOmni 2.0 Turbo) — quality and cost scale separately
- DreamTwin Character generates a styled AI persona from your reference photos (distinct from Face Swap)
Cons
- Credit drain is steep: 15 credits per 5-second SoulGen 3.0 clip means 100 monthly credits ≈ 33 seconds of video
- Free tier is thin — trial access only, watermarked output
- SoulOmni 2.0 Turbo (highest quality) costs 40 credits per 5-second 1080P clip — exhausts Pro credits in ≈12 seconds
SoulGen FAQ
Does SoulGen make realistic and anime NSFW images? Yes -- it has separate Realistic and Anime Character categories, with inpainting and outpainting on both.
What are the four video models on SoulGen and how do they differ? SoulGen 2.0 is the entry tier (10 cr/5s), SoulGen 3.0 is the current main model (15 cr/5s), SoulOmni 2.0 adds 720P/1080P with higher fidelity (18–30 cr/5s), and SoulOmni 2.0 Turbo is the top tier (25–40 cr/5s). All confirmed from the live pricing page. Higher tiers produce better motion and realism at significantly higher credit cost.
What is DreamTwin Character on SoulGen? DreamTwin Character generates a styled AI persona from your reference photos — distinct from Face Swap (which maps a face onto generated images). It appears in the image nav alongside Real Character and Anime Character as a separate creation mode.
Is SoulGen free? A free trial exists, but the free tier is limited and watermarked. Volume work requires a Pro plan (~100 credits/month).
How does SoulGen compare to Mage.space for images? SoulGen is more beginner-friendly with a cleaner UI and broader feature set (video, chat); Mage.space gives more fine-grained model and LoRA control for power users.
Scored on the public six-point method, assessed against documented features, pricing and real user reports (plus hands-on use where we have account access) · last updated October 2026.