Summary
This tutorial walks through Hicksfield’s newly released MCP (Model Context Protocol) server, a single-endpoint connector that gives AI agents — including Claude, Hermes, ChatGPT, Cursor, and Claude Code — the ability to generate videos, images, and clips directly without leaving the agent interface. The host demonstrates a sub-30-second setup process for both Claude (via the connectors UI) and Hermes (which is MCP-native), connecting to the endpoint at mcp.hixfeld.ai/mcp.
Once connected, the agent accesses Hicksfield’s full model stack: Seedens 2.5 for video generation and GPT Image 2 for images. Three concrete use cases are shown with real outputs and cost figures. A single product-shot video of a matte black water bottle on a concrete surface costs 45 credits (approximately $2.50) and completes in under a minute. A 15-second UGC-style ad built from the same product costs 75 credits (approximately $4) and runs in about four minutes from cold start — no product existed at the beginning of that timer. A multi-shot cartoon animation for a YouTube intro is generated through an iterative prompt-and-approve workflow, with the agent autonomously selecting the appropriate Hicksfield model based on the task description.
Additional tools in the MCP include a long-video clipper for vertical short creation, an aspect-ratio reframer, and a virality predictor. For creators and marketers running content pipelines, the integration collapses what previously required multiple separate platforms into a single agent prompt.
📺 Source: Sharbel A. · Published August 27, 2026
🏷️ Format: Tutorial Demo
