> ## Documentation Index
> Fetch the complete documentation index at: https://docs.vmeg.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Quick start

> Connect AI agents to VMEG media translation via Remote MCP

Connect VMEG to an MCP-compatible AI agent and translate video, audio, or subtitles through conversation — no custom integration code required. Usage draws on your existing VMEG plan credits.

<Info>
  **App marketplace / one-click connectors are not supported yet.** Use [manual setup](#add-vmeg-to-your-agent-manual) below with an [API Key](https://www.vmeg.ai/open-api-setting).
</Info>

<h2 id="add-vmeg-to-your-agent-manual">
  Add VMEG to your agent (manual)
</h2>

<Steps>
  <Step title="Get an API Key">
    Create a key on [API Configuration](https://www.vmeg.ai/open-api-setting). You need a paid plan with credits — see [Pricing](/guides/pricing).
  </Step>

  <Step title="Add a custom MCP server">
    In your [MCP-compatible agent](/mcp/quickstart#supported-products), open **Connectors** or **MCP servers** and choose **Add custom connector** (not a marketplace listing).
  </Step>

  <Step title="Enter the endpoint and Bearer token">
    * **Name:** `VMEG`
    * **URL:** `https://www.vmeg.ai/api/mcp`
    * **Authorization:** `Bearer YOUR_API_KEY`
  </Step>

  <Step title="Start translating">
    Ask your agent to translate media — for example, "Translate this video to Spanish with VMEG." The agent calls VMEG tools and bills against your plan.
  </Step>
</Steps>

Per-agent steps: [Cursor](/mcp/cursor), [Claude](/mcp/claude-web), and others in [Supported products](#supported-products). **ChatGPT** is [coming soon](/mcp/open-ai).

## What you can do

Once connected, your AI agent has access to the following tools:

### Materials

| Tool                   | Description                                                       |
| ---------------------- | ----------------------------------------------------------------- |
| `vmeg_list_materials`  | List uploaded video/audio materials in your project.              |
| `vmeg_delete_material` | Delete an uploaded material. Requires explicit user confirmation. |

### Voices

| Tool                     | Description                                                   |
| ------------------------ | ------------------------------------------------------------- |
| `vmeg_list_basic_voices` | List system voices. Filter by locale (e.g. `en-US`, `zh-CN`). |
| `vmeg_list_clone_voices` | List cloned voices in your project.                           |

### Create tasks

| Tool                                    | Description                                                                               |
| --------------------------------------- | ----------------------------------------------------------------------------------------- |
| `vmeg_create_video_translation_task`    | Create a video or audio translation task. Requires `targetLanguage` and one media source. |
| `vmeg_create_subtitle_translation_task` | Create a subtitle translation task. Requires `targetLanguage` and one media source.       |

### Tasks

| Tool               | Description                                          |
| ------------------ | ---------------------------------------------------- |
| `vmeg_list_tasks`  | List recent tasks or poll a single task by `taskId`. |
| `vmeg_delete_task` | Delete a task. Requires explicit user confirmation.  |

### Task draft (editor)

| Tool                        | Description                                              |
| --------------------------- | -------------------------------------------------------- |
| `vmeg_query_task_scripts`   | Query translated script lines.                           |
| `vmeg_query_task_subtitles` | Query subtitle clips.                                    |
| `vmeg_query_task_asr`       | Query ASR / original recognition lines.                  |
| `vmeg_query_task_speakers`  | Query speakers and voice assignments.                    |
| `vmeg_query_task_tracks`    | Query audio/video track blocks.                          |
| `vmeg_save_task_draft`      | Apply draft edits. Does not auto re-translate or export. |
| `vmeg_compose_task_draft`   | Render/export a new video or audio from the saved draft. |

### Async follow-up

| Tool                                     | Description                                    |
| ---------------------------------------- | ---------------------------------------------- |
| `vmeg_retranslate_task_scripts`          | Re-translate all scripts for an existing task. |
| `vmeg_get_retranslate_status`            | Poll re-translation status.                    |
| `vmeg_trigger_video_translation_all_tts` | Re-dub all or selected scripts.                |
| `vmeg_get_tts_status`                    | Poll per-script TTS dubbing status.            |

### Upload (HTTP)

| Endpoint                        | Description                                                                   |
| ------------------------------- | ----------------------------------------------------------------------------- |
| `POST /api/mcp/material/upload` | Multipart upload for local files. Returns `materialId` for create-task tools. |

<h2 id="media-sources">
  Media sources
</h2>

Each create-task call needs **exactly one** source:

| Source       | When to use                                           |
| ------------ | ----------------------------------------------------- |
| `materialId` | File already in your project (from upload or library) |
| `sourceUrl`  | HTTPS direct link — e.g. an agent attachment URL      |
| `youtubeUrl` | YouTube video URL only                                |

For `sourceUrl`, pass a public HTTPS file URL. For local files, upload via `POST /api/mcp/material/upload` first.

<h2 id="supported-products">
  Supported products
</h2>

| Agent                           | Manual custom MCP            | VMEG auth                                                     | Status          |
| ------------------------------- | ---------------------------- | ------------------------------------------------------------- | --------------- |
| [Cursor](/mcp/cursor)           | Yes — `mcp.json` or Settings | [API Key](https://www.vmeg.ai/open-api-setting) (Bearer)      | Available       |
| [Claude](/mcp/claude-web)       | Yes — Add custom connector   | OAuth on Connect                                              | Available       |
| [Claude Code](/mcp/claude-code) | Yes — `claude mcp add`       | [API Key](https://www.vmeg.ai/open-api-setting) or OAuth      | Available       |
| [Gemini CLI](/mcp/gemini-cli)   | Yes — `settings.json`        | [API Key](https://www.vmeg.ai/open-api-setting) (Bearer)      | Available       |
| [Manus](/mcp/manus)             | Yes — Custom MCP connector   | [API Key](https://www.vmeg.ai/open-api-setting) (Bearer only) | Available       |
| [ChatGPT](/mcp/open-ai)         | **No**                       | —                                                             | **Coming soon** |

Setup guides for each manual-capable agent are linked above. **App marketplace** one-click install is listed at the bottom of each guide as **coming soon**.

Any other MCP-compatible client can use the same endpoint:

```text theme={null}
https://www.vmeg.ai/api/mcp
```

## Remote MCP

|                | Remote MCP                                                       |
| -------------- | ---------------------------------------------------------------- |
| Setup          | Manual MCP URL + [API Key](https://www.vmeg.ai/open-api-setting) |
| Runs on        | VMEG hosted infrastructure                                       |
| Authentication | OAuth or [API Key](https://www.vmeg.ai/open-api-setting)         |
| Billing        | Your existing VMEG plan credits                                  |
| Best for       | Any MCP-compatible agent                                         |

<h2 id="app-marketplace">
  App marketplace
</h2>

<Note>
  One-click installs from marketplaces or app stores are **not supported yet**. **ChatGPT support is [coming soon](/mcp/open-ai).** For agents that support manual MCP, use [manual setup](#add-vmeg-to-your-agent-manual) with your [API Key](https://www.vmeg.ai/open-api-setting).
</Note>

## FAQ

<AccordionGroup>
  <Accordion title="Do I need an API Key?">
    For manual setup, yes — create an [API Key](https://www.vmeg.ai/open-api-setting) and send it as `Authorization: Bearer` on MCP requests. Some future marketplace connectors may use OAuth instead.
  </Accordion>

  <Accordion title="Does this cost extra?">
    No. Translation uses credits included in your VMEG plan. See [Pricing](/guides/pricing).
  </Accordion>

  <Accordion title="Which plans support MCP?">
    Free and paid VMEG accounts can connect MCP. Submitting translation tasks deducts credits from your account as usual — see [Pricing](/guides/pricing).
  </Accordion>

  <Accordion title="Can I use materials and voices from the vmeg.ai web app?">
    Yes. MCP binds to your **VMEG project** (via [API Key](https://www.vmeg.ai/open-api-setting) or OAuth) and uses the same materials, tasks, and project resources as the website.
  </Accordion>

  <Accordion title="Are MCP and Open API assets interchangeable?">
    No. Materials, tasks, and `voiceId` values created through the [Open API](/guides/introduction) are separate from MCP. Do not mix IDs across the two channels. See [API assets](/guides/assets/overview).
  </Accordion>

  <Accordion title="What's the difference between MCP and the Open API?">
    MCP exposes translation tools for AI agents (OAuth, VMEG project scope, shared with the web app). The Open API exposes REST endpoints for programmatic backends (API Key, separate asset store).
  </Accordion>
</AccordionGroup>