What it does
AgentHook exposes four tools for making short-form content with a reusable AI character. make_video takes a script and optional reference images and returns a talking-head video with native speech. make_image produces stills from a prompt or references. caption_video adds burned-in captions and a transcript to any video URL. create_influencer builds a reusable character, a portrait and character sheet, that later runs can reference with --influencer. The --enhance-prompt flag rewrites a rough one-liner into a more detailed prompt before generation, and --captions returns the video subtitled in tiktok or movie style.
It suits producing a series of videos or images that feature the same AI presenter. The character is created once, and later make_video and make_image runs can request it by name, so the same face appears across outputs.
It can be added as an MCP server, a skill or a CLI, and works in Claude Code, Cursor, OpenClaw, Hermes and n8n. All three options wrap the same API. New accounts receive 30 trial credits without a card. Failed runs are refunded automatically, and outputs are stored on permanent CDN URLs. The publisher provides a full API reference that agents can read before use. The CLI and skill are licensed under Apache-2.0.
Limitations and requirements
Generation draws on prepaid credits from an AgentHook account; there is no subscription. The hosted service is a commercial product.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
io.github.AvivK5498/agenthook
Character-consistent UGC video, images, and captions from any MCP client, via the AgentHook API.
v1.0.0As shown on registry.modelcontextprotocol.io when we last read it