CN8 API Documentation CN8 is an AI gateway: a single proxy layer that gives you access to AI capabilities — chatbots with RAG, multi-agent orchestration, LLM completions, dubbing, TTS, media generation, and production-grade video/audio processing — all through one consistent API with unified auth, billing, and usage tracking.
Documentation Services Platform API Chatbot └ Collections└ Chatbot CRUD└ Chat└ ConversationsAgent └ Agent CRUD└ Sub-Agents└ Sub-Agent Config└ Agent Chat└ Conversations└ User Context & Turn LogSearch └ Collections└ Search & HistoryVideo Chat └ Create & Manage└ Chat & Citations
Text to Speech └ Synthesize└ VoicesVoice Clone + TTS └ Persistent Clone└ Instant Clone + TTSAI Podcast └ Voices└ Plan└ SynthesizeMeeting Notes └ Submit & Poll└ Recording Layout
Dubbing └ Script└ RenderVoice Clone Dubbing
Avatar (Beta) Avatar (Cloud) └ Generate└ Portrait check
LLM Completions └ List & Compare Modelscn8 Architect └ Saved blueprints
Image from Text └ List Models└ Style Presets└ Generate ImageVideo from Text └ List Models└ Create└ Status & ResultVideo from Image └ List Models└ Create└ Status & ResultVideo Motion Control └ List Models└ Create└ Status & Result
Audio Description └ Script└ RenderTranscript └ Transcribe└ Metadata, Chapters & Quiz└ Scene Captions└ Translate└ Audio DescriptionDocument Infocard
Sentiment Analysis Named Entity Recognition Zero-Shot Classification Text Summarization Keyword Extraction Language Detection Question Answering Semantic Search Text Embeddings Text Similarity Few-Shot Classification
Background Remove / Replace Image Background Remove/Replace Video Upscale Image Upscale Smart Reframe Speech Clearance Audio Background Separation NSFW Detection Text Moderation Image Captioning Image Segmentation Image Inpainting Face & Privacy Blur (Image) Face & Privacy Blur (Video) Image Embedding Video Embedding
Services Media Generation Video from Image 64 / 109
Video from Image video-from-image
Asynchronous image-to-video generation — Veo, Seedance, Kling, Luma, Runway or MiniMax.
POST an image URL with generation_type: 'image-to-video', a required model (each model has its own per-second rate, from 4.6 to 156 credits per second depending on model and resolution) and an optional motion prompt. You get a job ID back; poll until the video is ready.
