Novita AI Platform Reference

Complete documentation for the Novita AI platform - an AI & Agent Cloud for developers.

When to Use This Skill

Load this skill when the user asks about:

• Novita AI products: Model APIs, GPU instances, serverless GPUs, agent sandbox
• Model information: "What models does Novita support?", model pricing, capabilities
• API guidance: How to use APIs, authentication, endpoints, parameters
• Pricing and billing: Cost estimates, billing queries, payment methods
• Integrations: LangChain, LlamaIndex, Cursor, and 30+ other tools
• Design system: Colors, typography, buttons, navigation, icons, logo
• Getting started: Quickstart guides, FAQs, setup instructions
• Troubleshooting: Error codes, common issues, support

Quick Reference

🔍 Quick: Query Available Models

Most common question: "What models does Novita support?"

Query Methods

1. Web Catalog (human-friendly):

• Browse 200+ models at https://novita.ai/models
• Filter by type: LLM, image, video, audio, embeddings

2. API Endpoint (automation):

curl https://api.novita.ai/openai/v1/models \
  -H "Authorization: Bearer <YOUR_API_KEY>"

Returns: Model ID, pricing per million tokens, context size, description

Model Categories

• LLMs: 100+ models (Llama, Qwen, DeepSeek, Mistral, etc.)
• Image Generation: Flux, Stable Diffusion, SDXL
• Video: Wan 2.6, CogVideoX
• Audio: TTS, voice cloning
• Embeddings: Text embedding models

Quick Links

Pro Tip: Always call /v1/models API first for the latest model list and current pricing.

How to Use This Documentation

1. Start Here

• New users: See getting-started/ - company overview, quickstart, FAQ
• Model queries: Check the "Quick: Query Available Models" section above
• API help: Jump to specific API reference sections below

2. Find Documentation by Category

Product Guides (usage and features):

• getting-started/ - Overview, quickstart, product pages
• llm/ - LLM API guides (16 files)
• model-apis/ - Model API guides (11 files)
• gpu-instance/ - GPU instances (14 files)
• serverless-gpus/ - Serverless GPUs (6 files)
• sandbox/ - Agent Sandbox (43 files)
• integrations/ - 30+ integration guides

API Reference (endpoints and parameters):

• api-reference/basic/ - Auth, billing (6 files)
• api-reference/llm/ - LLM endpoints (16 files)
• api-reference/image-apis/ - Image/video APIs (48 files)
• api-reference/gpu-instance/ - GPU APIs (2 files)

Support:

• billing/ - Billing and payments (4 files)
• team/ - Team management (1 file)

Design System:

• design-system/ - UI/UX specs (7 files)

3. File Naming Convention

Files are organized by category:

references/
├── getting-started/          # Product overviews and quickstart
├── llm/                      # LLM feature guides
├── model-apis/               # Model API guides
├── gpu-instance/             # GPU instance guides
├── serverless-gpus/          # Serverless GPU guides
├── sandbox/                  # Agent Sandbox docs (with subdirs)
├── integrations/             # Third-party tool integrations
├── api-reference/            # API endpoint documentation
│   ├── basic/               # Auth, billing APIs
│   ├── llm/                 # LLM API endpoints
│   ├── image-apis/          # Image/video API endpoints
│   └── gpu-instance/        # GPU instance APIs
├── billing/                  # Billing and payment
├── team/                     # Team management
└── design-system/            # UI/UX design specs

📚 Documentation Index

Core Product Documentation

Getting Started (8 files)

• company-overview.md - Company overview, products, testimonials
• gpus.md - GPU Cloud product overview
• sandbox.md - Agent Sandbox product overview
• gpu-baremetal.md - Bare metal GPU servers
• introduction.md - Platform introduction
• quickstart.md - Quick start guide
• faq.md - Frequently asked questions
• error-handling.md - Error handling

LLM Guides (17 files) Core: api · batch-api · function-calling · vision · reasoning · structured-outputs · prompt-cache · rate-limits · monitoring · observability-metrics · dedicated-endpoint · playgrounds · recommended

Model APIs (11 files) overview · sdks · dedicated-endpoints · training-guidance · custom-model · sampler · vae · clip-skip · rate-limits · v2-to-v3-migration · configure-custom-s3-bucket

GPU Instance (14 files) overview · overview-guide · choose-a-gpu · pricing · quickstart-* (5 files) · jupyterlab · save-image · upgrade-instance · edit-instance · image-prewarm

Serverless GPUs (6 files) overview · pricing · quickstart-* (4 files)

Agent Sandbox (43 files organized in subdirectories) Core: overview · pricing · sdk-and-cli

Quickstart: your-first-sandbox · introduction · installation · quick-start · frameworks · advanced

CLI: overview · auth · spawn · list · shutdown

Commands: overview · background · streaming

Filesystem: overview · read-write · upload · download · watch

Lifecycle: overview · clone · list · idle-timeout

Template: overview · customize-cpu-ram · start-cmd · ready-cmd · version-management

More: console · connect · internet-access · environment-variables · metadata · metrics · mount-cloudstorage

Integrations (30 tools) langchain · llamaindex · huggingface · cursor · dify · browseruse · skyvern · gradio · anythingllm · axolotl · chatbox · claude-code · codecompanion · continue · deepsearcher · docsgpt · helicone · kohya-ss-gui · langflow · langfuse · litellm · lobechat · lollms-webui · openai-agents-sdk · owl · pageassist · portkey · verba

API Reference

Basic APIs (6 files) authentication · error-code · get-user-balance · query-*-billing (3 files)

LLM APIs (16 files) list-models · retrieve-model · create-chat-completion · create-completion · create-embeddings · create-rerank · create-batch · cancel-batch · list-batches · retrieve-batch · list-files · upload-batch-input-file · query-file · retrieve-file-content · delete-file

Image/Video APIs (54 files) introduction

Core APIs: txt2img · img2img · inpainting · upscale · image-upscaler · remove-background · image-to-prompt · eraser · remove-text · replace-background · merge-face · reimagine · video-merge-face · task-result

Flux Models: flux-1-schnell · flux-1-kontext-dev · flux-1-kontext-max · flux-1-kontext-pro · flux-2-dev · flux-2-flex · flux-2-pro

Other Models: seedream-* (3) · glm-image · hunyuan-image-3 · qwen-* (2) · z-image-turbo · z-image-turbo-lora

Training: create-style-training · create-subject-training · list-training-task · get-training-images-url

Other: glm-tts-voice-clone · webhook

GPU Instance APIs (2 files) create-instance · list-clusters

Support & Design System

Billing (4 files) budgets · auto-top-up · payment-methods · low-balance-alert

Team (1 file) team-management

Design System (7 files) overview · typography · colors · buttons · navigation · icons · logo

Common Tasks

Start with Model APIs

1. Get API key from https://novita.ai/console
2. Set base URL to https://api.novita.ai/openai
3. Call /v1/models to list available models
4. Use OpenAI-compatible APIs for chat completions
5. See llm/api.md for details

Launch GPU Instance

1. Go to https://novita.ai/gpus-console/explore
2. Choose GPU or template
3. Configure and launch
4. Connect via SSH or web terminal
5. See gpu-instance/ for details

Create Serverless Endpoint

1. Prepare container image
2. Go to https://novita.ai/gpus-console/serverless
3. Create endpoint with scale policy
4. Test and deploy
5. See serverless-gpus/ for details

Start Agent Sandbox

1. Install SDK or CLI
2. Create sandbox with desired resources
3. Run commands or upload code
4. Pause/resume as needed
5. See sandbox/ for details

Integrate with Framework

1. Get Novita API key
2. Set base URL to https://api.novita.ai/openai
3. Update model names as needed
4. See integrations/ for specific guides

Support & Resources

• Documentation: https://novita.ai/docs
• Email: support@novita.ai
• Discord: https://discord.gg/YyPRAzwp7P
• FAQ: https://novita.ai/docs/guides/faq
• Status Page: https://status.novita.ai/