Novita AI
Access 200+ AI models through a unified API — LLM, image generation and editing, video generation, text-to-speech, speech recognition, and GPU cloud infrastructure.
- OpenAI-compatible LLM API works as a drop-in replacement with any OpenAI SDK
- 30+ image endpoints covering generation, editing, upscaling, background removal, face merging, and more
- Video generation from 7+ providers including Kling, Wan, Minimax Hailuo, Vidu, and Seedance
- Full GPU cloud management — instances, templates, storage, and serverless endpoints
Setup
- Get an API key at novita.ai/settings/key-management
- Set the environment variable:
export NOVITA_API_KEY=your_key - Base endpoint:
https://api.novita.ai
Services
| Service | Use When | Mode |
|---|---|---|
| LLM | Chat, completion, embeddings, reranking | Sync / Stream |
| Image Generation | Text-to-image (FLUX, SD, Seedream, Hunyuan, Qwen, GLM) | Sync / Async |
| Image Editing | Remove BG, upscale, inpaint, outpaint, cleanup, reimagine, merge face | Sync / Async |
| Video Generation | Text-to-video, image-to-video (Kling, Wan, Hailuo, Vidu, PixVerse, Seedance) | Async |
| Audio | TTS, ASR, voice cloning (MiniMax, GLM, Fish Audio) | Sync |
| Batch | Bulk LLM processing (OpenAI-compatible) | Async |
| GPU Cloud | Instances, templates, storage, serverless endpoints | Sync |
LLM (OpenAI-Compatible)
Drop-in replacement for the OpenAI API — use any OpenAI SDK with base https://api.novita.ai/openai.
import os
from openai import OpenAI
client = OpenAI(base_url="https://api.novita.ai/openai", api_key=os.environ["NOVITA_API_KEY"])
response = client.chat.completions.create(
model="moonshotai/kimi-k2.5",
messages=[{"role": "user", "content": "Hello"}],
max_tokens=512,
)Models: Kimi K2.5, MiniMax M2.7, GLM-5, DeepSeek V3, DeepSeek R1, and more via /openai/v1/models.
Features: vision (multimodal), reasoning, function calling, structured outputs, prompt caching, batch API.
Image Capabilities
| Feature | Description |
|---|---|
| Generation | FLUX.1 Schnell (fast, sync), FLUX Kontext, Stable Diffusion, Seedream, and more |
| Background | Remove background, replace with prompt-guided new background |
| Editing | Inpainting, outpainting, cleanup, reimagine, upscale |
| Face | Merge face from one image onto another |
| Analysis | Image-to-prompt — describe any image as text |
Video Capabilities
| Feature | Description |
|---|---|
| Text-to-video | Generate video from text via Kling, Wan, Hailuo, Vidu, Seedance |
| Image-to-video | Animate a still image with motion |
| Unified API | Single endpoint (/v3/video/create) for all video models |
Audio Capabilities
| Feature | Description |
|---|---|
| Text-to-speech | MiniMax (English, 17 voices, emotion control) and GLM (Chinese, low latency) |
| Speech-to-text | GLM ASR transcription |
| Voice cloning | Clone a voice from an audio sample |
GPU Cloud
Manage dedicated GPU instances, templates, network storage, and serverless endpoints for custom model deployment.
Security
- Never hardcode API keys — use environment variables or secret managers
- All media inputs should come from trusted, local sources only
- Enable NSFW detection for user-facing image applications
API References
For detailed endpoint parameters, request and response schemas, and code examples:
- LLM: references/llm-api.md — Chat, embeddings, rerank, function calling, structured outputs, batch
- Image: references/image-api.md — All generation and editing endpoints
- Video: references/video-api.md — Unified API and model-specific parameters
- Audio: references/audio-api.md — TTS variants, ASR, voice cloning
- GPU Cloud: references/gpu-api.md — Instances, templates, storage, serverless