CHANGELOG · CURRENT VERSION 1.6
What's new in Echoteka
Every release, in plain words. Echoteka is the online mastering lab for AI-made and home-recorded tracks.
Master a track free1 free master a day · no account
Echoteka 1.6 — November 9, 2026
- A new look. Tools and settings are compact tiles; fine tuning (sliders, the processing chain, presets) folds into a Pro panel. Four themes: Signal, Daylight, Soft and High contrast, plus System.
- Long tracks and audiobooks up to 2 hours. Up to 10 minutes costs 1 credit, every started 15 minutes adds 1 credit; the price is shown first, processing runs in the background, and credits come back if it fails. How it works.
- Audio to text. Transcripts with timecodes, TXT, DOCX, SRT and VTT, and a summary with topics and action items; 30 free minutes a day when signed in, then 1 credit per started hour.
- Refine with AI advice. On the result screen the AI turns its advice into settings for a new master, with the price shown before the run.
- Smart speech cleanup in the audio cutter and AI artifact cleanup with mastering in one run, with your own mastering settings.
1.5 — November 7, 2026
- An AI advice right after upload. The track studio suggests a style and a loudness with the reason in one sentence, and one button starts the run from that advice.
- A new mastering screen. It opens with “Auto will do: style and loudness”. Four sound variants appear on their own after the analysis with matched loudness, and a tap picks the style (signed-in users get 30 previews a day). “Soften harshness” and “Louder and denser” are one tap away; choose the format and 44.1, 48 or 96 kHz. How to choose a style before you master.
- Processing you can follow. Upload and processing are shown separately, with a time estimate, seven named steps, the track waveform and a “Finalizing” state on the last percent.
- A result that starts with a summary. What was done, the style, loudness before and after, the format and the sample rate; charts “Loudness over time” and “Spectrum before/after”; one “Original / Master” player with matched loudness and a loop.
- More air, less dullness. In the Universal and Trap styles the tonal correction now switches on only when a defect is found. Checked on 159 tracks; loudness and peaks do not change.
- Honest AI artifact cleanup. The cleanup says before the run that it weakens audible generator marks and hiss and does not promise that detectors will not find the track. If no marks are found, the token comes back. You can return to the task within 24 hours after closing the tab.
- Vocal removal results. “Download instrumental” and “Download vocal” on top, an “Original / Instrumental / Vocal” player, “Listen to 30 s” and “Tell me when it is ready” by email.
- Tools inside the app. Album mastering, stem mastering (2 to 8 stems), the mastering calculator, the loudness penalty, audio restoration with fine tuning, and a processing chain in “More settings”.
- Convenience. All formats the server understands are accepted; the app installs on a phone or a computer; password change, API keys, history export to CSV and account deletion; invite a friend and achievements; a weekly vote on what to build next for signed-in users.
1.4 — November 1, 2026
- Before and after in every result with matched loudness by default.
- Faster masters. The processing core is about 21% faster in our benchmark on reference tracks; the output was bit-identical in all 20 checks.
- Loud tracks stay loud. Without a loudness target of your own, a master is not turned down below −10 LUFS.
- Vintage sound, lo-fi and the music visualizer. Vintage sound, the lo-fi converter and the music visualizer.
- An assistant that knows the tools, and Spanish tool pages.
1.3 — October 24, 2026
- A second master that keeps the dynamics. If loudness has flattened a track, the result page offers a second version: a little quieter, with more dynamics kept. It is a separate file; the main master and its token charge do not change. How it works.
- AI label check. Track check has an AI label mode. It reads the file tags and the generator name when they are written, and gives a checklist with sources for Spotify, Deezer, DistroKid and Article 50 of the EU AI Act. It does not decide whether a track is AI-made, and it does not change the file.
- An honest cleanup result. If AI cleanup finds nothing to remove, the lab says so, the token comes back, and the daily limit is not used.
- Classical and Warm are gentler on dynamic tracks. On a very dynamic track, the limiter at the loudness target can squash it, so the target gives way a little, and the result says so. The louder version of the target sits right next to it.
- A Popular label. The tools people use most often in the lab, counted by runs over the last 30 days, carry a Popular label.
1.2 — October 10, 2026
- Hear it before you upload. A before/after player on a real Suno song on the home page and on Suno mastering — loudness-matched, and the switch happens at the same moment in the song. The LUFS checker has a “Try a 20-second sample” button, no file of your own needed.
- Master from your AI assistant. The MCP server at https://echoteka.ai/mcp works with Claude Code, Claude Desktop, Cursor and OpenAI Codex: analyze, clean AI artifacts, preview styles, master tracks and albums from a link. The free daily master applies; sign in from your assistant to use your account's bonus masters. There are no payment tools.
- Audio restoration. A new restoration page with sets for cassette, vinyl, phone and interview recordings. Hum removal now runs only when hum is found, a strength of zero really means off, and the report counts the clicks found and repaired.
- Clean and master in one run. An “Also clean AI artifacts” switch, free. The cleanup also downloads as MP3 or FLAC. Honest note: cleanup shifts pitch by 7 cents.
- Loudness follows your platform. The platform you choose sets the loudness; genre and advice change the sound, not the loudness. Speech presets and “Already loud” keep their own targets.
- More formats. Upload AAC, AIFF and WebM.
- An AI consultant on the site and in the Telegram bot @EchotekaAIBot, which also delivers your masters to Telegram. Invite a friend: +3 masters for you and +3 for them.
- New guides. Apple Music loudness, what AI artifact cleanup actually changes (2,230 tracks), Audiomack, Boomplay and JioSaavn loudness and a Suno export already at −13.6 LUFS.
- Sound fixes. No more bass dips on low 808 notes — the lowest-frequency filter was cutting twice as much as intended. Tracks that are already wide are no longer widened further, so they hold up better in mono.
1.1 — October 3, 2026
- A free master every day. One full master a day for everyone — with or without an account, days counted in UTC. Signing up adds 3 bonus masters. See free AI mastering.
- Free tools. A LUFS checker for loudness, true peak and loudness range, and a mastering calculator for platform targets.
- Suno loudness report. Original data from 248 Suno exports: how loud they are and how little they breathe. See the report.
- Honest comparisons. How Echoteka differs from LANDR, eMastered, BandLab, Masterchannel, Waves and Ozone — including when they are the better choice.
- Livelier sound at streaming loudness. At targets of −13 LUFS and quieter, multiband protection now follows the track's own loud passages instead of a fixed threshold, so masters keep more dynamics. Loud targets keep full protection.
- Breath control for voice. Loud breaths between phrases are brought down to a natural level, not cut out.
- Gentler AI artifact cleanup. Pitch is adjusted only when the spectrum needs it, and after a master you can clean the same track in one click.
1.0 — September 29, 2026
- Public launch. Online mastering for tracks made with Suno and other AI generators, and for home recordings.
- AI artifact cleanup for the metallic ringing and watery highs typical of AI tracks.
- Before/after report with loudness, true peak, loudness range, crest factor and stereo.
- 24 genre profiles, reference and album modes, streaming targets with a −1.0 dBTP ceiling.
- No subscription — 1 free master every day, with no account.
Echoteka versions its product on its own. The mastering engine underneath is shared and improved continuously; engine changes appear here when you can hear or see them.