io.github.AvivK5498/agenthook
Character-consistent UGC video, images, and captions from any MCP client, via the AgentHook API.
AgentHook
Your agent makes character-consistent UGC video.
Create one AI influencer, then reuse that same face in every video, image, and caption. Add it as an MCP server, a skill, or a CLI. Works in Claude Code, Cursor, OpenClaw, Hermes, and n8n.
Install
Pick one. All three wrap the same API, so none of them drift.
MCP add "npx -y @getagenthook/mcp" to your MCP client config
Skill npx skills add AvivK5498/agenthook --yes
CLI npm i -g getagenthook
Or paste this to your agent and let it set itself up:
Install the agenthook skill by running: npx skills add AvivK5498/agenthook --yes
Then follow the skill to set up AgentHook and make me a test image.
New accounts get 30 trial credits with no card. That covers a reusable influencer and a test image.
Agents: start here
Read https://getagenthook.com/llms-full.txt first. It is the full API contract, generated from the live schemas. Then:
npm i -g getagenthook # CLI (command: agenthook)
agenthook auth:login # prints a link + code for your human
agenthook run make_video --dry-run --prompt "..." # price it free, then drop --dry-run
What it does
| Tool | In, then out |
|---|---|
make_video | script + optional reference images, then a talking-head video with native speech |
make_image | prompt or references, then stills |
caption_video | any video URL, then burned-in captions + transcript |
create_influencer | brief idea + name, then a reusable character: portrait + character sheet, usable via --influencer on every later run |
The wedge is create_influencer. Make an influencer once, and your agent asks for that same face by name in every make_video and make_image run after it. One face, a whole series.
--enhance-prompt rewrites a rough one-liner into a director-grade prompt before generating, and --captions returns the video already subtitled in tiktok or movie style. Idea in, post-ready MP4 out, one command. Most platforms leave both steps to you.
Prepaid credits, no subscription. Failed runs refund automatically. Outputs live on permanent CDN URLs.
License
CLI and skill are Apache-2.0. The hosted service is a commercial product.
Installation
Source-derived launch command. Check the maintainer’s required arguments and credentials before running:
npx -y @getagenthook/mcpSet up in your AI client
Merge this template into ~/Library/Application Support/Claude/claude_desktop_config.json. Keep existing servers. Add any arguments, credentials, and permissions required by the maintainer; this template has not been install-tested.
{
"mcpServers": {
"io-github-avivk5498-agenthook": {
"command": "npx",
"args": [
"-y",
"@getagenthook/mcp"
]
}
}
}Restart Claude Desktop completely for changes to take effect. Confirm the server appears connected in the client’s tool list, then try a read-only example from its documentation.
Claude Desktop setup referencePackage
@getagenthook/mcpnpmCompatible MCP Clients
io.github.AvivK5498/agenthook works with any MCP-compatible client. Copy the config snippet from the Configuration section above and add it to the file shown for your client, then restart the application.
- Claude Desktop
~/Library/Application Support/Claude/claude_desktop_config.jsonRestart Claude Desktop completely for changes to take effect. - Cursor
~/.cursor/mcp.jsonRestart Cursor for changes to take effect. - VS Code
.vscode/mcp.jsonReload VS Code window for changes to take effect. - Windsurf
~/.codeium/windsurf/mcp_config.jsonRestart Windsurf for changes to take effect. - Claude Code
.mcp.jsonSave at the project root, then start Claude Code in that project and review the MCP server approval prompt. Keep real credentials out of shared files.