Aura AI已验证已精选
@auraai.app
关于 Aura AI
Access 20+ AI video and image models from Google, OpenAI, xAI, ByteDance, and more. Generate and edit videos with Veo 3.1, Sora 2, Seedance, Kling O3, and images with FLUX, Nano Banana, Seedram 4, GPT Image — all in one
概览
What is Aura AI MCP
Aura AI exposes its full media-generation catalog as a Model Context Protocol server, so any MCP-compatible client (Claude Desktop, Claude Code, custom agents) can generate and edit images and videos directly, using the same models available on auraai.app.
Tools
- list_capabilities — compact overview of every model available (image, image-edit, text-to-video, image-to-video, reference-to-video)
- get_models — full detail for a single category: supported ratios, durations, resolutions, per-second pricing, audio support
- generate_image — text-to-image across 15+ models (FLUX, Imagen 4, Ideogram V4 Instant / Fast, Nano Banana 2, Grok, GPT Image 2, Seedream 5, MAI Image 2.5, and more)
- generate_video — text-to-video and image-to-video across 30+ models (Veo 3.1, Sora 2 Pro, Kling O3 4K, Seedance 2.0, Wan 2.7, HappyHorse, LTX 2.3, Luma Ray 2, Pixverse 5, Grok Video, and more)
- edit_image — natural-language image editing with Seedream 5 Pro, Nano Banana 2 Edit, GPT Image 2 Edit, Grok Edit, Wan 2.7 Edit
- reference_to_video — generate a video from 1–5 reference images (Kling O3 Reference, Veo 3.1 Reference, Seedance 2.0 Reference, Vidu Q2, Wan 2.7 Reference)
- blog_to_video — one-shot: paste a blog post URL, get back a vertical UGC-style video with voiceover, b-roll and captions
Authentication
OAuth 2.0 with PKCE. The MCP server delegates authentication to auraai.app — users approve access in their browser, tokens are minted and stored client-side. No API keys to manage.
Requirements
- An Aura AI account (free signup at auraai.app)
- Image / video credits on the account (Starter plan and up for premium models and reference-to-video / blog-to-video)
Endpoint
https://auraai.app/mcp (Streamable HTTP + SSE for long-running generations)
Model selection
Every generative tool accepts selected_model: "auto" (default), which routes to the best model for the requested output and the user's plan. You can also pass an explicit model value from list_capabilities to pin the choice.
Docs
Full protocol docs and per-tool schemas at auraai.app/mcp.
基本信息
分类
其他
作者
auraai.app
评论