MCP Server

An AI other AIs can use. Connect any MCP-capable agent to Scarleta with a key and it calls our audio tools natively, as real tool-calls. No hand-rolled REST plumbing. Give any agent a full audio toolbelt.

Native tool-calls

Exposes the same audio capabilities as the REST and batch API over the Model Context Protocol, so any MCP-capable AI can call them directly.

Connect with a key

Authenticate the connection with your customer API key, the same key you use everywhere else on the platform.

Billed per use

Charged per use on the same per-second-of-audio meter as the rest of the platform. Nothing new to learn, same prepaid tokens.

Connect your agent

Point your MCP client at Scarleta's server and pass your API key. Once connected, the audio tools appear to your agent alongside its other tools.

{
  "mcpServers": {
    "scarleta": {
      "url": "https://api.scarleta.ai/mcp",
      "headers": {
        "Authorization": "Bearer $SCARLETA_API_KEY"
      }
    }
  }
}

A sample of the toolbelt

The MCP server exposes Scarleta's full set of audio capabilities. A few of them:

stem_analysis

Split a song into separate stems.

remove_audio_element

Remove or isolate one element of a track.

musical_intelligence_analysis

Detect key and chord progression.

tempo_analysis

Detect BPM, beat times, and steadiness.

speech_analysis

Transcribe spoken audio.

noise_reduction

Clean up unwanted noise.

Same capabilities as RESTLive

Next steps

Chat / Agent API Batch API Pricing