Underlying AI Models
Last Updated: 2025-07-18
Deep Nostalgia AI uses mainstream, industry-recognized generative AI models supplied by their original developers. We do not relabel, rebrand, or misrepresent these models. Every model used in the Service retains its original developer's safety guardrails. This page is provided to comply with our payment partners' AIGC transparency requirements and to help our users understand which underlying models power our image and video generation features.
Image Generation Models
| Model Name | Original Developer | Use Case |
|---|---|---|
| Flux.1 Kontext / Flux 2 Max | Black Forest Labs | Text-to-image, image-to-image editing |
| Seedream 4.0 / 5.0 Pro | ByteDance | Text-to-image, high-fidelity image synthesis |
| Nano Banana (Gemini 2.5 Flash Image) | Text-to-image, image-to-image editing | |
| Nano Banana 2 / 2 Lite / Pro | Text-to-image, image-to-image editing | |
| GPT Image 1.5 / GPT Image 2 | OpenAI | Text-to-image, image-to-image editing |
| Z-Image Turbo | Zhipu AI (Z.ai) | Ultra-fast text-to-image generation |
| Hunyuan Image 3.0 | Tencent | Text-to-image, photorealistic synthesis |
Video Generation Models
| Model Name | Original Developer | Use Case |
|---|---|---|
| Seedance 1.0 Lite / Pro / 1.5 Pro / 2.0 / 2.1 | ByteDance | Image-to-video, text-to-video |
| Kling 1.5 / 2.0 / 2.5 / 2.6 / O1 | Kuaishou | Image-to-video, text-to-video |
| Veo 3 (Standard / Fast) | Text-to-video with native audio | |
| Wan 2.1 / 2.2b / 2.5 / 2.7 / Animate / Effects | Alibaba Cloud | Image-to-video, text-to-video, video effects |
| Hailuo 02 / 2.3 Fast | MiniMax | Image-to-video, text-to-video |
| Sora 2 (Standard / HD) | OpenAI | Text-to-video, image-to-video |
| Vidu 1.5 / 2.0 / Q1 | Shengshu Technology | Image-to-video, text-to-video |
| Pixverse (Standard / V6 / Effects) | Pixverse (Aihug Tech) | Image-to-video, video effects |
| Hunyuan Video | Tencent | Text-to-video, image-to-video |
| Gemini Omni | Multimodal conversational video |
Content Safety Mechanisms
Every underlying model used by the Service retains the safety guardrails implemented by its original developer. In addition, the Service applies its own pre-generation prompt moderation layer (a local banned-keyword library and a semantic classifier) and configures the upstream model provider's built-in safety checker (enable_safety_checker=true and nsfw_checker=true) on every generation request. For the full description of our content moderation pipeline, please see our Acceptable Use Policy.
Trademarks
All model names, brand names, and trademarks listed on this page are the property of their respective owners. Their appearance on this page is for transparency and disclosure purposes only and does not imply any affiliation, endorsement, or sponsorship.
Contact
For questions about this AI model disclosure page, please contact us at [email protected].