Quickstart

Connect an AI assistant to Ghostfeed and create your first avatar in about two minutes.

Ghostfeed's agent platform lets AI assistants create and manage content in your Ghostfeed account, through an MCP server and a REST API that expose the same operations, accept the same OAuth tokens, and answer in the same response envelope.

Connect

One server, one URL, OAuth in the browser:

https://api.ghostfeed.ai/api/v2/mcp

The fastest path: paste this to your agent and let it set itself up.

Paste into your agent

Set up Ghostfeed for me so I can create AI avatars, UGC videos, and
slideshows from here.

1. Add the MCP server (Streamable HTTP):
   https://api.ghostfeed.ai/api/v2/mcp
   In Claude Code: claude mcp add --transport http ghostfeed https://api.ghostfeed.ai/api/v2/mcp
2. Complete the OAuth sign-in it opens (Google).
3. Install the companion skills: npx skills add ghostfeed-ai/skills --skill '*'

Once that's done, let me know when it's ready.

Streamable HTTP. Sign-in is Google; the consent screen picks which workspaces the connection may act in. Tokens are stored by your client and refresh automatically.

Claude Code

Add the server:

Terminal

claude mcp add --transport http ghostfeed https://api.ghostfeed.ai/api/v2/mcp

Run /mcp in a session, pick ghostfeed, choose Authenticate.

Claude apps

Desktop, web, and mobile all read the same connector list, so this is a one-time setup.

Open Settings → Connectors, click Add → Add custom connector.

Name it Ghostfeed and paste the remote MCP server URL (leave the OAuth fields empty; sign-in handles it):

https://api.ghostfeed.ai/api/v2/mcp

Click Add, then Connect, and finish the sign-in in your browser.

Claude renders Ghostfeed's interactive widgets here: the avatar builder, crop editor, media galleries, and results all appear as live UI in the chat.

Codex CLI

Add the server:

Terminal

codex mcp add ghostfeed --url https://api.ghostfeed.ai/api/v2/mcp

Authenticate:

Terminal

codex mcp login ghostfeed

ChatGPT

Open Settings → Plugins.

Click Add → Add MCP server, name it ghostfeed, choose Streamable HTTP, and paste the server URL:

https://api.ghostfeed.ai/api/v2/mcp

Save, then complete the sign-in when the first Ghostfeed tool runs.

On older ChatGPT builds the same setup lives under Settings → Apps → Create with developer mode enabled.

Grok

Open grok.com/connectors, click New Connector, choose Custom.

Paste the server URL:

https://api.ghostfeed.ai/api/v2/mcp

Complete the sign-in.

Grok Build

Add the server:

Terminal

grok mcp add --transport http ghostfeed https://api.ghostfeed.ai/api/v2/mcp

The sign-in opens in your browser the first time a tool runs; or run /mcps in the TUI and press i to authenticate.

Cursor & VS Code

Both editors use a small JSON file; only the file name and top-level key differ.

Cursor: add the server to ~/.cursor/mcp.json (global) or .cursor/mcp.json (project):

.cursor/mcp.json

{
  "mcpServers": {
    "ghostfeed": { "url": "https://api.ghostfeed.ai/api/v2/mcp" }
  }
}

VS Code: add the server to .vscode/mcp.json (or run MCP: Add Server from the Command Palette and choose HTTP):

.vscode/mcp.json

{
  "servers": {
    "ghostfeed": { "type": "http", "url": "https://api.ghostfeed.ai/api/v2/mcp" }
  }
}

Complete the sign-in when the first Ghostfeed tool runs.

Anything else

Any MCP client that speaks Streamable HTTP and OAuth 2.1 works: an unauthenticated request returns 401 with discovery metadata, the client registers itself (Dynamic Client Registration), and PKCE completes in the browser. There is no SSE endpoint and no static API keys.

Skills

To get the best out of the Ghostfeed MCP and 10x the quality of your harness, pair it with our skills.

Install the complete Ghostfeed skill set:

Terminal

npx skills add ghostfeed-ai/skills --skill '*'

Pick your agents and scope at the prompt. The wildcard selects every Ghostfeed skill in the package. Global installs land in the shared ~/.agents/skills/ directory, which Claude Code, Codex, Cursor, Grok Build, and the rest of the skills CLI's 70+ agents read.

No Node? Download the complete skill bundle:

Terminal

curl -LsS https://api.github.com/repos/ghostfeed-ai/skills/tarball/main \
  -o ghostfeed-skills.tar.gz

Browse skills under Skills.

On this page