Frequently Asked

TrackSensei is an AI production mentor for electronic music producers. You upload an audio bounce of your track, optionally together with your Ableton project (.als), and TrackSensei analyzes them to give you a detailed production report, scored against the conventions of your genre. Then you can chat with an AI mentor that knows your specific track, ask about your mix, your arrangement, your sound design, anything you'd ask a more experienced producer sitting next to you.

TrackSensei is built for electronic music producers at any level, from bedroom producers working on their first release to semi-pros polishing a track for a label. If you make electronic music and want direct, detailed feedback on your production, this is for you. Every analysis starts from an audio bounce, from any DAW. Ableton users get the deepest analysis, because we also read the .als project file.

Project file analysis (full project structure, plugin chains, MIDI, routing) is currently Ableton-only via .als files. But you can upload your audio bounce from any DAW (FL Studio, Logic Pro, Cubase, Bitwig, Reaper, and more) in WAV, MP3, FLAC, AIFF or OGG and get a full audio analysis with mix feedback, frequency analysis, loudness, stereo balance, and AI chat. Native FL Studio and Logic Pro project file support is on the roadmap. Join the waitlist below to be notified when your DAW is supported.

12 main genres, 71 subgenre profiles, one measurement engine. Techno and house targets come straight from 563 released tracks; the other genres are built from in-depth genre research and get sharper as more tracks come in. The main genres are Techno, House, Trance, Drum & Bass, Dubstep, Hip-Hop, Hardstyle, Hardcore, Pop, Rock, Ambient and Breaks. Each profile has its own LUFS, frequency balance, stereo and arrangement targets, so Industrial Techno gets different benchmarks than Minimal Techno, Gabber gets different benchmarks than Frenchcore, and Tech House gets different benchmarks than Deep House.

ChatGPT can talk about music in general but it has never heard your track. TrackSensei actually analyzes your audio: it measures your loudness, frequency balance, stereo image, dynamics, groove and arrangement, then scores mix balance, dynamics and groove against targets specific to your genre. Arrangement gets written feedback instead of a score. The AI then interprets those numbers in the context of your specific track and gives you concrete, actionable feedback. ChatGPT can give you general advice about mixing techno. TrackSensei can tell you what is actually going on in your techno track right now.

Audio: WAV, MP3, FLAC, AIFF, OGG. Project files: Ableton Live .als. Audio is required for every analysis; the .als file is optional and gives the deepest feedback, so TrackSensei works best with Ableton. Both are read in your browser (audio through the Web Audio API), so nothing is uploaded to a server: file size is bounded by browser memory, and files up to about 500 MB typically work smoothly.

A typical analysis takes under two minutes, depending on the length and complexity of your track. You'll see a live progress indicator while it's processing.

The analysis is built on industry-standard signal processing: ITU-R BS.1770 LUFS metering, True Peak, PLR/PSR dynamics, 24-band Bark-weighted spectrum, per-band stereo correlation, and MFCC-based section detection. Scores are calculated against genre-specific targets. 12 main genres, 71 subgenre profiles, one measurement engine. Techno and house targets come straight from 563 released tracks; the other genres are built from in-depth genre research and get sharper as more tracks come in. The AI feedback builds on this measurement data and is generally very useful, but remember: it is a tool, not a replacement for your ears or your creative direction. Use it as a knowledgeable second opinion.

Every analysis is free. You get production scores, how you compare to the reference for your genre, every issue we found with its severity and measured value, and the full fix for your top issue. In a project you can also compare your latest version with the one before it. Free covers 2 analyses and 3 chat questions a month. What you buy on top of that is the explanation behind every other issue, Deep Dives and compares with older versions.

Free: scores, reference comparison, every issue with its severity and measured value, the full fix for your top issue and a compare of your latest two versions, 2 analyses and 3 chat questions a month. Release Check (€5.95, one track): every fix explained, Deep Dives, issue threads, 20 mentor-chat messages on our best model and one extra version, open for 14 days, with the report readable for a year. Pro (€12/month): the full report on every track you upload (every fix explained, deep dives, version compare) and 150,000 feedback tokens a month, about 15 full track journeys. Ultimate (€29/month): 400,000 tokens (about 40 journeys), longer and more detailed mentor answers, Stem Interaction Analysis and Boosters.

One payment of €5.95 that unlocks the full report for one track: every fix explained with why it matters, how to fix it and the expected effect, pillar Deep Dives, issue threads, mentor chat on our best model (20 messages) and one extra version to check your fix, including the compare view. No subscription, no renewal.

A Release Check is €5.95 and covers one track. Pro is €12 a month and gives you the full report on every track you upload, with a monthly token budget (about 15 full track journeys) instead of a message count. If you finish more than one track a month, Pro is the cheaper door.

Chat, Deep Dives, issue threads, version compare and the extra version close after 14 days. The report itself stays readable for a year, so every explanation you paid for stays where you left it. You can unlock chat and new versions again for €5.95, or go Pro.

Yes. A Release Check includes one extra version of the same track within 14 days, with the compare view that shows which change actually moved your scores. On Free you get 2 analyses a month and every new version uses one of them. On Pro and Ultimate, versions run on your monthly token budget.

A Release Check is digital content delivered straight away, so your right of withdrawal ends the moment you pay. The Stripe checkout page states that before you confirm. If something went wrong on our side, mail us and we refund it.

Yes. Once a track is unlocked, the results page has a download button that gives you the full report as a PDF: the scores, every issue with why it matters, how to fix it and the expected effect, and your Deep Dives. The feedback in it is in the same language as your analysis. A Release Check keeps that download open for a year; on Pro and Ultimate it stays open while you are subscribed.

Yes. You can cancel your subscription at any time from your dashboard. Your access continues until the end of your current billing period, no prorated refunds, but no surprise charges either.

Yes, Stripe automatically generates and emails invoices for every payment. If you're a business customer and need invoices with your company name and VAT number, you can add your billing details in the customer portal. Prices include VAT where it applies; Stripe works out the VAT share from your billing details at checkout and shows it on your invoice.

We accept all major credit and debit cards (Visa, Mastercard, American Express) as well as iDEAL, SEPA Direct Debit, and other local payment methods through Stripe.

TrackSensei detects your location and displays prices in your local currency where supported. Today that's EUR (€12/mo Pro, €29/mo Ultimate) for the eurozone and GBP (£10/mo Pro, £25/mo Ultimate) for the UK; we'll add USD and others as we expand. You can override the auto-detected currency from the switcher in the top bar at any time. Stripe Checkout charges you in the same currency you saw on the site, so there's no surprise on your bank statement. VAT is calculated automatically based on your billing address. Visitors from countries without a hand-picked price still see EUR on the site, and Stripe converts to the local currency at checkout using the mid-market rate.

Every AI request consumes a number of output tokens. A Deep Dive, a chat reply, a stems analysis: each one draws from the same shared pot. Pro comes with 150,000 tokens a month and Ultimate with 400,000, resetting on your signup date. Free and the Release Check are not token-budgeted: Free gives 3 chat questions a month, a Release Check gives 20 messages on one track.

Rough costs per action: 1 Deep Dive (the full analysis report) is about 3,000 tokens. 1 chat message is about 1,200 tokens (a focused follow-up answer). 1 Stems Interaction Analysis is about 3,000 tokens. So a typical track journey of 1 Deep Dive plus 6 chats lands around 10,000 tokens. Pro covers about 15 of those per month (or 50+ quick analyses), Ultimate about 40 (or 130+ quick analyses). The token budget is a subscription thing: Free and the Release Check are capped by count instead, analyses and chat questions on Free and messages on a Release Check.

Boosters are one-time purchases of additional feedback tokens that never expire. Three sizes: Booster S (250k, €15), M (500k, €30), L (1M, €60). They're consumed only after your monthly Ultimate budget runs out, so you always burn the cheaper subscription rate first. Per-token, Boosters are about 17% cheaper than the Ultimate subscription rate.

Boosters are an Ultimate-only feature. The intent is that a hobbyist who hits the budget once or twice a month upgrades to Ultimate first; Boosters cover the higher-volume Ultimate users who occasionally need extra headroom for a release week. If you're on Pro and frequently run out, upgrading to Ultimate gives you about 2.5× the budget for €17/month more.

No, never. Once added to your account they stay there until used. You can buy a Booster today and have the tokens still available three months from now. They're consumed only when your monthly budget is empty, and only by the amount each request actually costs.

We absorb minor pricing fluctuations as a cost of business, so your budget and Booster tokens stay the same. For a structural model-price change we'd update the monthly budget per tier (with at least 30 days' notice to active subscribers, per the Terms) and recalibrate Booster token counts at the next batch. We won't retroactively change the size of a Booster you've already purchased.

Your audio is analyzed entirely in your browser (Web Audio API) and is never uploaded to our servers. Your Ableton .als project is read in your browser as well, and only a structural summary of it is sent. What reaches us are measurements such as LUFS, spectral balance and correlations. Analyses in a project are kept while you work on it; after 120 days without a new version the full reports and chats are deleted, and analyses outside a project are deleted after a year. A report unlocked with a Release Check stays readable for a year after purchase. For the written feedback we use Anthropic's Claude API, which doesn't use API data for model training. The privacy policy lists every retention period and every service provider we use.

No. We use Anthropic's Claude API, which explicitly does not train on API data. Your music, projects, and conversations are not used to improve any AI model. Your work stays yours.

Yes. We use industry-standard security: encrypted connections (HTTPS), hashed passwords, rate limiting to prevent abuse, and strict input validation on all endpoints. Payment data is handled entirely by Stripe, we never see your card number.