Skip to content
audo

Audio & Video MCP Server: Claude Code, Cursor, ChatGPT & Claude

Add Audo once, and your AI assistant or coding agent can run all 37 audio and video tools plus multi-step workflows for you. The remote Streamable HTTP MCP server is at:

https://audo.ai/mcp

It signs you in through your browser, with OAuth. Clients that can't sign in can send an API key from the API page as a bearer token.

Setup for each assistant has the steps, with one-click buttons for Claude, Cursor, and VS Code.

How files get in

  1. In the chat. Claude and Cowork show an upload box in the conversation. In ChatGPT on the web or desktop, attach the file to your message.
  2. From your computer. Coding agents such as Claude Code, Codex, Cursor, and Gemini CLI upload the file themselves with curl, and save each result next to the original. Allow audo.ai and files.audo.ai once so they don't ask every time.
  3. With an upload link. Any assistant can give you a link to an upload page. Open it and drop the file: no sign-in, any size, and it carries on if the connection drops. Or upload on this site, and your assistant finds it in your files.

add_file takes a link or a ChatGPT attachment. Otherwise it returns the upload page link for the person, and, for agents with a terminal, a signed upload link with ready-to-run curl commands. Every result comes with download links, and a curl command that saves it next to the original. Review each transfer URL and keep normal client approvals. Never authorize curl by command prefix or by matching audo.ai in command text.

What your assistant can do

  • Audio and video tools (37): remove_noise, enhance_voice, split_audio, transcribe_audio, align_script, cut_audio, remove_silence, adjust_loudness, convert_audio, join_audio, replace_audio, analyze_audio, remove_fillers, translate_transcript, translate_subtitles, find_highlights, shorten_transcript, level_speakers, mix_tracks, change_speed_pitch, find_key_bpm, reverse_audio, mute_video, convert_video, compress_video, crop_video, resize_video, reverse_video, loop_video, get_frame, video_to_gif, add_captions, audio_to_video, add_overlay, reframe_plan, reframe_video, text_to_speech — including speech noise removal (remove_noise), echo removal and voice restoration (enhance_voice), 4-stem music and dialogue separation (split_audio), 60-language transcription with automatic speaker diarization (transcribe_audio), filler-word and pause cutting (remove_fillers, remove_silence), active-speaker 9:16 video reframing (reframe_video), styled burned captions (add_captions), and text-to-speech (text_to_speech).
  • Multi-step workflows: list_workflows, run_workflow, get_workflow_run, and save_workflow. Run built-in pipelines (podcast_ready, fix_video_sound, clips, meeting_shortener) or chain up to 30 steps in one server-side WorkflowRecipe without round-tripping intermediate files.
  • Presets, files, and jobs: list_caption_presets, save_caption_style, list_audiogram_presets, get_account, add_file, list_files, get_file, get_job, cancel_job, and get_transcript.
  • A tool or workflow waits up to 40 seconds and hands back the finished result with download links; longer runs return a job_id or run_id to check with get_job or get_workflow_run.
  • Pricing is pay-as-you-go (pay once from $5, with no monthly subscription and credits that never expire), charged only if the job succeeds. When the balance is short, the free allowance runs a few times a day, just like on the website.

Example prompts

  • Remove the background noise from interview.wav, then transcribe it.
  • Remove the long pauses from episode-12.mp3 and make it podcast-loud.
  • Make SRT captions for launch-video.mp4.
  • Pull the audio out of this video as an MP3.

Quick setup for coding and terminal agents

Setup for each assistant has step-by-step instructions for all 13 supported assistants (ChatGPT, Claude, Cowork, Claude Code, Codex, Cursor, Gemini CLI, VS Code, Copilot CLI, Antigravity, Zed, OpenCode, and Raycast). For terminal agents, one command adds Audo:

Terminal
# Claude Code:
claude mcp add --transport http --scope user audo https://audo.ai/mcp

# OpenAI Codex:
codex mcp add audo --url https://audo.ai/mcp

# Gemini CLI:
gemini mcp add --transport http --scope user audo https://audo.ai/mcp

# GitHub Copilot CLI:
copilot mcp add --transport http audo https://audo.ai/mcp

In OpenCode, add "audo": { "type": "remote", "url": "https://audo.ai/mcp", "enabled": true } under mcp in ~/.config/opencode/opencode.json and run opencode mcp auth audo. In Cursor and VS Code, use the one-click install buttons on /setup.

This page as Markdown