Gives Claude access to 45+ image models, 70+ video models, and 20+ audio tools through the kie.ai API. Every model includes embedded research from Averiguare with prompt techniques, cost analysis, and competitor comparisons. You can filter models by capability ("lip sync", "cheapest video", "text rendering") and the server searches across tags and research fields. Handles generation across OpenAI, Google Veo, Flux, Sora, Kling, Alibaba's HappyHorse and Wan, Suno, ElevenLabs, and dozens more. Includes utilities for extending videos, upscaling to 4K, vocal separation, and TTS. Tracks pricing in credits and USD so you can optimize for cost. Runs as stdio for Claude Desktop or HTTP for remote deployments.
claude mcp add --transport stdio elibarnett-kie-mcp uvx kie-mcp