Skip to main content
Glama
README.md
# Importly MCP Server

Turn any video or audio link into **files, transcripts and metadata**, right from your AI agent: YouTube, TikTok, X, podcasts, and 1000+ other sites.

[Importly](https://www.importly.io/?utm_source=mcp&utm_medium=integration&utm_campaign=readme) is one API for your app, and the same key works in your agent over MCP. It works from agent sandboxes and cloud servers where downloads usually get blocked, so you never deal with blocked requests, broken downloaders, or videos without captions.

## Tools

| Tool | What it does |
| ---- | ------------ |
| `import_media` / `check_import` | Download media to hosted storage, get a direct link (optionally with a transcript) |
| `transcribe_media` / `check_transcription` | Speech-to-text with timestamped segments |
| `get_metadata` / `check_metadata` | Full metadata: formats, uploader, thumbnails, stats |
| `get_basic_metadata` / `check_basic_metadata` | Fast title/duration/thumbnail lookup |
| `check_balance` | Remaining credit balance |
| `get_integration_guide` | Working code (Node, Next.js, Python, cURL) to add Importly to your own app: key setup, webhooks, polling, errors. No credits used |

Jobs are async: submit tools return a `jobId`, check tools poll it. Responses tell the agent exactly what to do next.

Transcripts use the video's own captions when it has them ($0.01). For media without captions, the agent can retry with `provider: "importly"` for AI speech-to-text ($0.02/min).

## Setup (hosted, recommended)

Importly runs this server for you at **`https://mcp.importly.io/mcp`** (Streamable HTTP). Nothing to install.

### Claude.ai, ChatGPT, Claude Desktop

Add a **custom connector** with the URL `https://mcp.importly.io/mcp`, click **Connect**, and sign in to Importly (new accounts get **$3.00 in free credits, no card required**). No API key needed.

### Coding agents (API key)

1. Get an API key at [importly.io](https://importly.io/?utm_source=mcp&utm_medium=integration&utm_campaign=readme) — **$3.00 in free credits, no card required**.
2. Add the server to your MCP client, sending the key as an `Authorization: Bearer` header:

#### Claude Code

```bash
claude mcp add --transport http importly https://mcp.importly.io/mcp --header "Authorization: Bearer your_key"
```

#### Cursor

```json
{
  "mcpServers": {
    "importly": {
      "url": "https://mcp.importly.io/mcp",
      "headers": { "Authorization": "Bearer your_key" }
    }
  }
}
```

More clients (VS Code, Cline, …): [setup guide](https://importly.io/docs/integrations/mcp?utm_source=mcp&utm_medium=integration&utm_campaign=readme).

> **This npm package is no longer maintained.** The hosted server above runs the same tools and always has the latest version. The source here is kept for reference.

## Example prompts

- "Transcribe this talk and summarize the key points: https://www.youtube.com/watch?v=…"
- "Download the audio of this podcast episode and give me a link"
- "How long is this video and who uploaded it?"
- "Add import-from-URL to our upload form using Importly" (the agent calls `get_integration_guide` and writes the code, with the same API key read from your environment)

## Pricing

Usage-based credits (downloads ~$0.01/MB, transcription per media minute) — see [pricing](https://importly.io/pricing?utm_source=mcp&utm_medium=integration&utm_campaign=readme). New accounts get $3.00 free.

## Docs & support

- [Setup guide](https://importly.io/docs/integrations/mcp?utm_source=mcp&utm_medium=integration&utm_campaign=readme)
- [API docs](https://importly.io/docs?utm_source=mcp&utm_medium=integration&utm_campaign=readme)
- support@importly.io

MIT licensed. The server is a thin wrapper over the [Importly API](https://importly.io/docs?utm_source=mcp&utm_medium=integration&utm_campaign=readme).

TDQS

A4.2/5.0

Scored across 9 tools

Disambiguation4/5

Tools are clearly organized into four async operations (import, full metadata, basic metadata, transcription) each with a paired check tool, plus a balance check. The only ambiguity is between import_media with includeTranscript and transcribe_media, but descriptions clarify the intended use.

Naming Consistency5/5

All tool names follow a verb_noun pattern with action verbs (import_media, get_metadata, transcribe_media) and check_ prefixed status tools (check_import, check_metadata, check_basic_metadata, check_transcription). The pattern is consistent and predictable.

Tool Count5/5

With 9 tools, the server is well-scoped for its purpose: four async operations, four corresponding status checks, and a balance check. Each tool serves a distinct need without redundancy.

Completeness4/5

The server covers the core workflows of media import, metadata retrieval (full and basic), transcription, and balance monitoring. Minor gaps exist such as no job cancellation or listing, but these are not essential for the stated purpose.

Maintenance

ActivityMaintained
ResponsivenessLimited data