# eonik [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/eonik-ai/eonik-mcp  
**GitHub Stars:** 1  
**npm Downloads (last month):** 194  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/eonik

## Description
Read brand, research, and performance facts; draft briefs. Never spends. You approve every cut.

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

```json
"mcpServers": {
  "eonik": {
    "command": "npx",
    "args": ["-y","eonik-mcp"],
    "env": {
      "EONIK_API_KEY": "",
      "EONIK_API_URL": "",
      "EONIK_BRAND_ID": ""
    }
  }
}
```

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

## Documentation

## What eonik MCP server does

The eonik MCP server gives an MCP-compatible assistant read access to advertising context stored in an eonik account. It can retrieve brand information, notes saved for later, and ads that have been saved in the account. It can also surface competitor ads already archived by eonik, including information about what ran, its duration, and how the ad was constructed.

The assistant can read Meta performance data as factual reporting, including spend, click-through rate, and status. These values are presented as facts rather than forecasts. Public ads can be examined scene by scene, giving the assistant material to use when helping prepare an advertising brief.

The connection is intentionally read-and-draft oriented. The assistant does not pause, scale, stop, launch, or upload campaigns, and it does not write to Meta, TikTok, or Google Ads. Timeline editing and export remain in the eonik Mac application, while publishing requires a human action in Ads Manager.

## How it works

An MCP host starts the package with `npx` and passes an eonik API key in the process environment. The assistant then uses the connected account as its source for brand, research, saved-ad, competitor, and performance information. The resulting workflow is suited to reviewing evidence, identifying a useful starting ad, and producing a brief for human approval.

The Mac application remains the place where the actual cut is made. It can build from existing footage and use AI to fill gaps, but those editing actions are outside the MCP connection. The eonik MCP server does not claim that an ad will work; it supplies account data and analysis for a person to evaluate.

## Setup and configuration

Create an API key in eonik workspace settings, then add the server to the MCP host configuration. A typical stdio entry is:

```json
{
  "mcpServers": {
    "eonik": {
      "command": "npx",
      "args": ["-y", "eonik-mcp"],
      "env": {
        "EONIK_API_KEY": "eonk_live_..."
      }
    }
  }
}
```

`EONIK_API_KEY` is required. `EONIK_API_URL` is optional and defaults to `https://api.eonik.ai`. Set `EONIK_BRAND_ID` when the account contains more than one brand. Keep the API key in the MCP host environment and do not commit it to a repository.

You can also run a local build by installing dependencies, building the project, setting the API key, and starting `build/index.js`:

```bash
npm install && npm run build
export EONIK_API_KEY=eonk_live_...
node build/index.js
```

## Tools and capabilities

The eonik MCP server supports these grounded tasks:

- Read brand details, remembered notes, and saved ads.
- Inspect archived competitor ads and their observed construction.
- Retrieve Meta spend, CTR, and status as reported facts.
- Analyze a public advertisement scene by scene.
- Provide source material for drafting an advertising brief.

Claude Desktop, Cursor, and Codex are named as supported connection targets in the project documentation.

## Limitations and notes

The connection cannot modify advertising platforms or make budget decisions. It cannot launch, pause, scale, or terminate ads, and it cannot write to Meta, TikTok, or Google Ads. Uploading an ad is a manual action in Ads Manager.

The server also does not replace the Mac application for timeline edits or export. The documented desktop requirement for that application is macOS 15 or later on Apple silicon. Performance information should be treated as historical or current facts supplied by the account, not as a prediction of future results.

_Full upstream README: https://allmcps.com/mcp/eonik/readme_