MCP App Store
by KeenEthics
WaveLocus icon

WaveLocus MCP Server for ChatGPT

By SAHIL P KARAWADE

Overview

Check short voice and narration recordings locally in your browser, then discuss the measurements in ChatGPT. Review up to three independent mono or stereo excerpts, each up to 60 seconds and 16 MiB. WAV is supported. MP3, M4A/AAC and FLAC depend on browser decoding support. WaveLocus measures sample peak, whole-excerpt RMS, DC offset and full-scale samples. Analysis v2 also estimates true peak and measures gated loudness of the selected excerpt in LUFS when available. Unavailable measurements include a reason instead of a guessed value. You can select a background-only passage to measure its unweighted level, but WaveLocus does not verify that speech is absent or measure equipment noise floor. Compare an original with a revised copy of the same passage, or approve an optional bounded gain-adjusted WAV when eligible. Your original files are unchanged. Audio and filenames stay on your device. Connecting alone shares no report. Only selected structured results are sent after separate approval. Shared reports expire within 24 hours. This does not remove noise or echo, transcribe speech, master music, judge audio quality or certify publishing compliance. The narration-level reference is partial and conditional, not a universal voice target. True peak is an estimate and excerpt LUFS does not describe an unseen complete programme.

How to connect

  1. ChatGPT: Open the WaveLocus listing in the ChatGPT app directory and choose Connect. ChatGPT listing WaveLocus exposes 2 tools here.

Tools

get_shared_recording_review

ChatGPT
WaveLocus computes the measurements locally in the user's browser. The editor field is user-selected editing context only. Do not say the recording was measured, analysed, checked or verified in Audacity or another editor. Measurements do not prove that an editor was used or which editing operation occurred. Retrieves only a privately approved report. Each request consumes persisted request and session-operation allowances, but never edits, deletes or reshares the report. Do not poll or retry repeatedly. Read explicitly shared v1 or v2 reports. For v1 preserve channelReferenceFacts and nextStep. For v2 preserve sampleReferenceFacts and sampleReferenceNextStep. Explain intended use, editor, measurement availability and method limits. True peak is an estimate. LUFS describes this excerpt, not an unseen programme. Background-only selection is user-declared, not verified noise detection or SNR. Do not invent delivery targets, infer quality, prescribe processing from new metrics or treat unavailable values as zero. Independent reference testing is not conformance certification. Batch items are independent excerpts, not a ranking. No audio, filename, locator or local revision-factor note is returned. Earlier chat replies do not update after a new share.

start_recording_review

ChatGPT
Start a browser-local check of one to three WAV, MP3, M4A/AAC or FLAC excerpts, where the browser supports decoding. Measures sample peak and whole-excerpt RMS, with optional bounded gain adjustment and separate per-recording findings. V2 also returns estimated true peak and gated excerpt LUFS when available. It does not certify conformance or publishing compliance. Creates one temporary connection-bound handoff. No raw recording is uploaded. For an original and an edited copy of the same passage, add only the original in Step 1. In Step 3, open "Check a file revised in your editor", select the revised copy and the changed factor, confirm the same passage, selection boundaries and channel order, then choose "Check revision". In Step 4, separately approve sharing the single before/after report. Do not instruct the user to add both copies as independent Step 1 files or share both reports. Independent batch slots are not a controlled before/after comparison. Measurement differences do not prove what processing caused the change or whether quality improved. Use recording-review tools only when the user explicitly requests voice or narration measurements, an optional volume adjustment, a same-passage comparison, or retrieval of an explicitly shared report. This configuration supports sample peak, whole-excerpt RMS, estimated true peak and gated excerpt LUFS when available. Independent reference testing is not conformance certification or a publishing approval. For a request only to remove hiss, noise or echo, transcribe speech, master music or certify publishing compliance, do not call any WaveLocus tool or create a review session. Explain the unsupported capability directly. A WaveLocus @mention is not a measurement request. For mixed requests, explain the unsupported part and perform only the explicitly requested supported check. If intent is unclear, clarify before creating a session. Never promise noise removal, transcription or mastering from measurements. A user-selected background passage is not verified noise detection, equipment noise floor or SNR. All tool requests consume stored request allowances. Approved report reads also consume a finite session operation allowance without changing the report. Do not poll or retry repeatedly.

Capabilities

Writes

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/wavelocus"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/wavelocus)