EggAPI

One API key and a remote MCP that lets Claude Code and other MCP clients generate images and videos with Seedream, Nano Banana, GPT Image, Seedance and Wan.

Hosted MCP Server

npx add-mcp 'https://api.eggapi.ai/mcp'

Installs into Claude Code, Codex, Cursor and more

Documentation

EggAPI runs a remote Model Context Protocol server. Any MCP client that supports Streamable HTTP can list models, start image and video tasks, and read their results with your API key. Usage is billed to the key's account exactly like POST /v1/generate.

SettingValue
Endpointhttps://api.eggapi.ai/mcp
TransportStreamable HTTP
Auth headerAuthorization: Bearer YOUR_API_KEY

Create a key in the dashboard and export it first:

export EGGAPI_API_KEY="egg_..."

Client Setup

Claude Code

claude mcp add --transport http eggapi https://api.eggapi.ai/mcp \
  --header "Authorization: Bearer $EGGAPI_API_KEY"

Codex

Add to ~/.codex/config.toml:

[mcp_servers.eggapi]
url = "https://api.eggapi.ai/mcp"
bearer_token_env_var = "EGGAPI_API_KEY"

Cursor and other clients

Add to your client's mcp.json:

{
  "mcpServers": {
    "eggapi": {
      "url": "https://api.eggapi.ai/mcp",
      "headers": { "Authorization": "Bearer <EGGAPI_API_KEY>" }
    }
  }
}

Step-by-step setup with a key generator is also on eggapi.ai/connect.

ChatGPT

First create an API key in the EggAPI Dashboard. You need a ChatGPT account that can use MCP, plugins, or connectors.

Use https://api.eggapi.ai/mcp/k/<YOUR_API_KEY>, replacing <YOUR_API_KEY> with your key. The URL path carries authentication; no extra Authorization header is needed. This URL is equivalent to your secret key. Do not share it.

Tested macOS desktop path — ChatGPT 26.1103.41825:

Add to ~/.codex/config.toml:

[mcp_servers.eggapi]
url = "https://api.eggapi.ai/mcp/k/<YOUR_API_KEY>"

In ChatGPT, open Settings → Plugins → MCPs and enable eggapi. The first tool call asks “Allow ChatGPT to use eggapi?”; allow it. Image and video generation worked in this test.

Official web alternative — docs verified 2026-11-10; not re-tested by us:

OpenAI's Connecting custom MCP servers in ChatGPT describes this current web flow. Developer Mode is not required first.

  1. Open Plugins, select the plus button, then Add custom MCP server. Enter a name, such as eggapi, and an optional description.
  2. Under Connection, enter the public HTTPS MCP URL above. Choose No authentication if asked: the URL already carries your EggAPI key.
  3. Review the risk warning and check I understand and want to continue. Select Create as a plugin, then review the discovered tools.
  4. Install the plugin, start a new conversation, type @, and select the plugin.

The developer guides Add custom MCP server and Connect and test your plugin use the same Plugins flow. Supported transports are SSE and streaming HTTP; documented authentication options are OAuth, No authentication, and OAuth or no authentication.

For this web alternative, the Help page documents custom MCP connections and full MCP support for Business and Enterprise/Edu. Its FAQ says Pro custom MCP connections have read/fetch permissions, and custom MCP servers cannot be connected on mobile (web only).

Example prompts:

  • “Use eggapi with GPT Image 2 (openai/gpt-image-2/text-to-image) to generate a product photo of a ceramic mug on a white background.”
  • “Use eggapi with Wan 3.0 (alibaba/wan-3.0/text-to-video) to generate a 5-second video of steam rising from a coffee cup.”

Tools

ToolArgumentsReturns
list_modelskind? (image or video)Model IDs with kind, generation_type, unit_price_usd, unit_price_by_resolution, and the tool that runs each one.
generate_imagemodel, prompt, images?, negative_prompt?, aspect_ratio?, resolution?, num_outputs?, parameters?, nsfw_filter?A task_id. images holds the source image URLs for edit models.
generate_videomodel, prompt, duration, image?, negative_prompt?, aspect_ratio?, resolution?, parameters?, nsfw_filter?A task_id. image is the first frame URL for image-to-video models; duration is required.
get_tasktask_idstatus (pending, processing, completed, failed), done, output urls, cost, and error.

The agent's workflow is:

  1. Call list_models to pick a valid model ID.
  2. Call generate_image or generate_video. It returns a task_id immediately; the output is not ready yet.
  3. Poll get_task every 5 to 15 seconds until done is true. Images usually finish within a minute; videos usually take one to three minutes.

Tasks created over MCP also appear in GET /v1/tasks and in your usage records.

Dynamic UI tools

The same server also provides get_ui_catalog, render_ui, update_ui, get_ui_state, set_ui_values, and run_ui_action: the agent composes a form page from a short JSON description, the user fills it in inside the conversation (clients that support MCP Apps) or in a browser, and the agent reads the result to continue. See Dynamic UI.

Workflow Skills

For complete workflows on top of these tools, such as e-commerce image sets, ad creatives, and short films, install the ready-made workflow skills.