# Corent [Health: Active]

**Category:** 🎥 Multimedia Process  
**Repository:** https://github.com/gg13121/corent-mcp  
**GitHub Stars:** 0  
**npm Downloads (last month):** 283  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/corent-2

## Description
Generate images and videos via the Corent API — auto model routing with provider fallback.

## Tools
Capabilities this server exposes over MCP:

- **plan** — Describe a request in plain language; Corent returns the plan (image vs video, tier, settings) + cost estimate, without generating
- **create** — Describe what you want + a spend ceiling; Corent decides everything and generates it (the "zero decisions" path)
- **generate_image** — Text → image, synchronous, returns a permanent URL
- **generate_video** — Text (or image) → video, async job
- **get_job** — Poll a job until completed
- **get_balance** — Remaining account balance
- **get_status** — Live tier health

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "corent": {
    "command": "npx",
    "args": ["-y","corent-mcp"],
    "env": {
      "CORENT_API_KEY": ""
    }
  }
}
```

**Requires environment variables:** `CORENT_API_KEY` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What Corent MCP server does

The Corent MCP server gives an MCP-compatible AI agent access to image and video generation through the Corent API. An agent can describe the desired result in natural language instead of selecting a model or provider directly. Corent handles the routing decision, including whether the request is for an image or video and which tier and settings to use.

The `plan` tool provides a proposed media type, tier, settings, and cost estimate without starting generation. The `create` tool follows the same agent-oriented approach but generates the requested result and accepts a per-call spending ceiling. This makes it suitable for autonomous workflows that need a budget boundary.

## How it works

Image generation is synchronous: `generate_image` accepts a text prompt and returns a permanent URL. Video generation is asynchronous: `generate_video` accepts text or an image and returns a job that can be checked with `get_job` until it completes.

Corent also exposes account and service information through `get_balance`, which reports the remaining account balance, and `get_status`, which reports live tier health. Automatic model routing and provider fallback are handled by the Corent API rather than by the agent.

## Setup and configuration

A Corent API key is required. Create one through corent.tech, then provide it as the `CORENT_API_KEY` environment variable when starting the local server. The package can be launched through `npx -y corent-mcp` in an MCP client configuration.

For local development, install the npm dependencies, build the project, and start `dist/index.js` with `CORENT_API_KEY` set. Claude Code can add the server with its MCP command and the same environment variable.

A hosted Streamable HTTP endpoint is also available at `https://mcp.corent.tech/mcp`. Send the key in an `Authorization: Bearer ...` header; `X-Corent-Api-Key` is also accepted. The hosted endpoint supports query-parameter authentication for Smithery compatibility, but the documented guidance is to prefer headers because URLs may be recorded in logs.

## Tools and capabilities

- `plan`: Estimates a request and proposes image or video output, tier, and settings without generating it.
- `create`: Generates from a natural-language request while enforcing a supplied spend ceiling.
- `generate_image`: Produces an image synchronously and returns a permanent URL.
- `generate_video`: Starts an asynchronous video job from text or an image.
- `get_job`: Checks the state of a video job.
- `get_balance`: Reports the remaining account balance.
- `get_status`: Reports current tier health.

## Limitations and notes

The Corent MCP server requires an active Corent API key and an available account balance. Video requests require asynchronous job handling, so an agent must use `get_job` to monitor completion. The `plan` tool estimates and describes a request but does not generate media. Hosted deployments should keep API keys in headers rather than query strings when possible.

_Full upstream README: https://allmcps.com/mcp/corent-2/readme_