EchoJay

EchoJay is an AI-powered mix feedback tool that combines professional-grade audio metering with a language model that reads those measurements and tells you what to do about them. The product exists to close a specific gap: general-purpose AI assistants can listen to a track and describe it, but they do not measure it. There is no LUFS reading, no true peak value, no crest factor, just impressions. EchoJay replaces impressions with hard numbers taken from your actual session, and then explains what those numbers mean for the genre you are working in.

The metering suite is the foundation. Loudness metering implements EBU R128 with K-weighted filtering and proper gating, reporting momentary, short-term, and integrated LUFS plus loudness range, using full track averages rather than snapshots. True peak detection runs at four times oversampling and sits alongside RMS, peak, crest factor, and DC offset, so headroom before a bounce is unambiguous. Stereo analysis covers real-time width, correlation, vectorscope, and mid-side balance, catching phase problems and mono compatibility issues before a listener hears them. An eight-band average spectrum is rendered as a smooth EQ curve, giving an at-a-glance view of tonal balance.

The AI layer is what makes those readings actionable. The model receives every meter value with each review and gives specific advice tailored to the selected genre, from hip-hop, drill, and trap to pop, R&B, house, rock, jazz, classical, lo-fi, afrobeat, phonk, and dozens more across thirty-plus supported genres. You can ask follow-up questions, request a plugin chain with exact settings, or go as deep as you want. Stem analysis shifts the focus depending on what you select: sibilance and presence for vocals, transient punch and low-end weight for drums, mono compatibility for bass. More than sixty channel types are supported, and channel detection flags a mismatch if the meter data does not match the selected channel type, so feedback is never wasted on the wrong signal.

Several release-oriented features extend the workflow. Reference tracks can be dropped in and analysed locally, storing LUFS, EQ curve, and stereo image so you can compare your mix against commercial releases. AI Compare puts your mixes and references side by side, including full albums of up to eight tracks, and returns a detailed breakdown covering consistency, loudness spread, and tonal balance. Revision tracking detects a V2 automatically, keeps the chat together, and reports exactly what changed. There is a personalised setup for your monitoring, room treatment, and plugin list, plus client mix notes so briefs are factored in, and an audio-reactive particle visualiser with multiple presets and colour themes.

EchoJay runs in two places. The web app processes audio in the browser with the Web Audio API, and Pro users can optionally store mixes for playback on any device. The plugin is the same engine inside your session, available as VST3, AU, or AAX on macOS and Windows, working on any channel or bus. Audio never leaves your machine in plugin mode; only meter values are sent to the AI.