Best MCP servers for audio and transcription
7 working options, checked Nov 8, 2026.
Short answer: the top picks for audio and transcription right now are Transcribe, voicememotext and Arcmira: YouTube Transcript Search. We list 7 options here: 2 work without sign-in and 5 need a sign-in.
Audio MCP servers turn speech into text, make music and sound effects, or search what was said in videos. They are useful for meeting notes, subtitles and any project where an assistant needs to work with sound.
Top picks
-
Transcribe
Pay-per-minute speech-to-text for AI agents: MP3/WAV in, transcript, timestamps and SRT/VTT out.
Tools include
transcribe_audio -
voicememotext
Turn voice memos, meetings and any audio URL into text or SRT from your MCP client.
-
Arcmira: YouTube Transcript Search
Find who said what with timestamps, discover topics across videos and livestreams, and distinguish organic recommendations from sponsored ad reads.
-
Neume
Make songs, change one part's lyrics in the same voice, and reimagine songs in new styles.
-
SFXMint
CC0 sound effects for games and videos, with a free JSON API and an MCP server.
Tools include
search_sounds,get_role_sound,get_sound,get_sound_set
All 7 compared
| Server | What it does | Setup | Tools |
|---|---|---|---|
| Transcribe | Pay-per-minute speech-to-text for AI agents: MP3/WAV in, transcript, timestamps and SRT/VTT out. | Hosted, no sign-in | 1 |
| voicememotext | Turn voice memos, meetings and any audio URL into text or SRT from your MCP client. | Hosted, sign-in | – |
| Arcmira: YouTube Transcript Search | Find who said what with timestamps, discover topics across videos and livestreams, and distinguish organic recommendations from sponsored ad reads. | Hosted, sign-in | – |
| Neume | Make songs, change one part's lyrics in the same voice, and reimagine songs in new styles. | Hosted, sign-in | – |
| SFXMint | CC0 sound effects for games and videos, with a free JSON API and an MCP server. | Hosted, no sign-in | 4 |
| Seedeo MCP | Generate AI videos, images and music from Claude or Cursor with Veo 3.1, Kling, Seedance and Suno. | Hosted, sign-in | – |
| Reclip | Video, audio and image tools for apps and AI agents | Hosted, sign-in | – |
How to choose
- Transcribe or create? Speech-to-text servers read audio. Music and sound-effect servers make it.
- Need timestamps? For subtitles, choose a server that outputs SRT or VTT files.
- Check the license. For generated audio in commercial work, confirm your rights on the maker's site.
How to connect one
Here is Transcribe as an example. In Claude Code, run:
claude mcp add --transport http transcribe https://api.youraitools.co/transcribe/mcpFor other apps, follow our guides for Claude, ChatGPT, Cursor and VS Code.
How we picked these
We read each listing's tools and put the closest fit for this job first. Hosted servers must pass a real MCP handshake in our daily check to stay on this page, and local servers are reviewed from their source or package. That shows a server responds, not how well it works for you, and we don't run every tool in production.
Frequently asked questions
Which of these MCP servers work without signing in?
Transcribe and SFXMint answered our latest live check with no account or key.
Can an AI assistant transcribe audio through MCP?
Yes. Give a transcription server an audio file or URL and it returns the text, often with timestamps or subtitle files.
Can MCP search inside YouTube videos?
Yes. Transcript servers let an assistant find who said what, with timestamps, without watching the video.
How often is this list updated?
Every hosted server here is re-checked daily with a real MCP handshake, and a server drops off this page while it fails that check. Last check: Nov 8, 2026.