Skip to main content

Release Notes

Release Notes

September 14, 2024

🚀 FLUX.1 [dev] ControlNet - Text to Image is now available on Novita AI.

Integrate them via API: .

Currently, we support the following ControlNet models:

  • XLabs-AI/flux-controlnet-depth-v3.safetensors (Preprocessor: DepthAnythingPreprocessor)
  • XLabs-AI/flux-controlnet-canny-v3.safetensors (Preprocessor: CannyEdgePreprocessor)
  • XLabs-AI/flux-controlnet-hed-v3.safetensors (Preprocessor: HEDPreprocessor)

September 9, 2024

🔥 FLUX.1 [dev] LoRA - Text to Image is now available on Novita AI.

You can integrate them via API: .

Currently, we support the following LoRA models:

  • AWPortraitFL_LoRA_1.0.safetensors,
  • Disney-Studios-Flux-000008.safetensors
  • F1_Q_Lora.safetensors
  • F1_natural_illustration.safetensors
  • Studio_Ghibli_Flux.safetensors
  • flux_realism_lora.safetensors
  • healing_illustration.safetensors

September 6, 2024

📢 Video Merge Face API is now available on Novita AI.

You can integrate them via API: .

August 30, 2024

🔥 FLUX.1 models Image-to-Image APIs are now available on Novita AI.

You can integrate them via API:

  • FLUX.1 [dev]:
  • FLUX.1 [schnell]:

August 26, 2024

🎙️ FaceFusion model is now available on Novita AI, which is the next-generation face swapper and enhancer.

More details: .

August 15, 2024

🔥 FLUX.1 models are now available on Novita AI.

You can integrate them via API:

July 24, 2024

🔥 Llama 3.1 models are now available on Novita AI.

You can try them out in the playground:

Or integrate with API: .

July 19, 2024

Novita AI has released "GPU Instance", which provides cost-effective, accessible high-performance computing with easy GPU access, ideal for complex AI tasks. Users benefit from on-demand rental, optimizing efficiency and cost-effectiveness.
For more details, please visit: https://novita.ai/gpu-instance.

June 20, 2024

Now you can use Stable Diffusion 3 Medium model with the Text-to-Image API and the Image-to-Image API.
For more details, please visit: .

June 2, 2024

For the asynchronous V3 APIs, to receive the generated outputs, in addition to polling the Task Result API,
now you can also listen to "ASYNC_TASK_RESULT" event on your webhook endpoint.

Also, you can test your integration without making actual API calls by enabling Test Mode, which will help you thoroughly test your logic.
More details can be found at: .

May 30, 2024

Novita AI has open-sourced a new Video-Generation model on GitHub: https://github.com/novitalabs/AnimateAnyone, which is the unofficial implementation of AnimateAnyone.
You try it out in the Novita AI Playground: https://novita.ai/model-api/playground#animate-anyone.

May 28, 2024

The img2img Playground now supports ControlNet and IPAdapter,
visit https://novita.ai/model-api/playground#img2img to have a try.

And txt2img Playground allows you to share your parameters with a share-link.
Here is my share for you: https://novita.ai/model-api/playground?share=2930971716886178.

May 21, 2024

Novita AI has released the Relight API, which helps you enhance or change the illumination of images and seamlessly replace the background.
And you can utilize all SD1.x models deployed by Novita AI in the Relight API.
👉 .

May 9, 2024

Now Novita AI has released the Text-To-Speech API and Voice Clone Instant API.

Text-To-Speech API generates realistic speech in various languages
👉 .
Voice Clone Instant API clones your voice with only a few minutes of audio
👉 .

April 29, 2024

Now you can chat with your favorite character by using the novita.ai LLM features.
Play it now 👉 https://novita.ai//model-api/llm-api/playground

April 19, 2024

🔥 Llama 3 models are now available on novita.ai.

You can try them out in the playground:

April 11, 2024

We've recently added more hot LLM models for you, now the supported models include:

  • lzlv_70b (https://huggingface.co/lizpreciatior/lzlv_70b_fp16_hf)
  • Nous-Hermes-2-Mixtral-8x7B-DPO (https://huggingface.co/NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO)
  • gryphe/mythomax-l2-13b (https://huggingface.co/Gryphe/MythoMax-L2-13b)
  • nousresearch/nous-hermes-llama2-13b (https://huggingface.co/NousResearch/Nous-Hermes-Llama2-13b)

April 8, 2024

Now novita.ai allows you to search for the models you like, making it easier to find the model you need.
Please feel free to give it a try and enjoy the experience: https://novita.ai/model-api/model.

April 7, 2024

Now the LLM Chat API supports seamless integration into your product, offering compatibility with the OpenAI API standard. 🚀
More details: .

March 21, 2024

InstantID now supports the ControlNet lineart model (controlnet-softedge-sdxl-1.0).
You can try it in the playground: https://huggingface.co/spaces/novita-ai/InstantID.

March 20, 2024

We've released novita.ai Python SDK version 0.5.0 🐍

This version introduces the txt2img_v3 and img2img_v3 methods for calling the API v3 txt2img and img2img,
and you can find examples here: https://github.com/novitalabs/python-sdk/tree/main/examples.

March 14, 2024

We've released novita.ai Javascript SDK version 1.0.35, available at https://www.npmjs.com/package/novita-sdk/v/1.0.35.

This version introduces the img2ImgV3 and txt2ImgV3 methods for calling the API v3 txt2img and img2img,
and you can find examples here: https://github.com/novitalabs/javascript-sdk/blob/main/examples.

February 29, 2024

We've launched some new features lately, here are the details:

  1. You can now check if the civitai models exist on novita.ai. Simply copy the civitai model page URL and paste it into the search box at https://novita.ai/model-api/model.
  2. The InstantID API has been launched. Please refer to the API documentation at: https://docs.novita.ai/case-api/instant-id, or you can access the InstantID playground at https://huggingface.co/spaces/novita-ai/InstantID.
  3. SDXL Lightning LoRAs are now supported in the API /v3/async/txt2img. Available models include sdxl_lightning_2step_lora, sdxl_lightning_4step_lora, and sdxl_lightning_8step_lora.

View More