MCPthe API built for agent CLIs
Let Claude Code and Codex
generate images and video.
EggAPI turns Seedream, Nano Banana, GPT Image 2, Seedance and Wan into MCP tools your agent can call. Add it from the terminal with one command, then just ask: the agent picks the model, runs the task and fetches the file. One key, one bill.
works withClaude CodeCodexCursorany MCP client
claude · example session
$claude mcp add --transport http eggapi https://api.eggapi.ai/mcp ✓ Added HTTP MCP server eggapi ❯Make a 5-second product video for the landing page with Seedance 2.5, black background. ●eggapi.generate_video(model: "bytedance/seedance-2.5/text-to-video", duration: 5) ⎿ task queued · video usually takes 1–3 min●eggapi.get_task(task_id) ⎿ completed · output.mp4 ⏺Done. The video is ready, link below.
connect
One command, and your agent can create.
Remote MCP over HTTP, authenticated with your EggAPI key. No SDK to install, no local process to run.
terminal
claude mcp add --transport http eggapi https://api.eggapi.ai/mcp \ --header "Authorization: Bearer $EGGAPI_API_KEY"
Export EGGAPI_API_KEY in your shell first, then run the command. Type /mcp inside Claude Code to check the connection.
how it works
From one sentence to a finished file, inside the conversation.
01
Create a key
Sign in, create an API key in the dashboard and add credits.
02
Connect your agent
Add EggAPI as an MCP server in Claude Code, Codex or any MCP client.
03
Just ask
The agent picks a model, starts the task, polls it and hands you the file URL.
list_models
(kind?)
List the models your key can use, with type and price, so the agent can pick within budget.
generate_image
(model, prompt, images?, aspect_ratio?, resolution?, num_outputs?)
Text-to-image, or edit with reference images.
generate_video
(model, prompt, duration, image?, aspect_ratio?, resolution?)
Text-to-video or image-to-video. Returns a task id right away.
get_task
(task_id)
Check task status and fetch the output URLs.
models
Six models. Nothing to sift through.
A short list of image and video models worth using, each with its newest versions.
Give your agent a camera.
Create a key, paste one command into your terminal, start generating.