# Intrect — full answer corpus > Every question Intrect answers in public, with the page it comes from. The curated index is at https://intrect.io/llms.txt; this file is the answers themselves, at https://intrect.io/llms-full.txt. Generated from each page's own FAQPage JSON-LD by `scripts/build_llms_full.py`. Do not hand-edit — edit the page and rerun. 81 answers across 16 pages. Citation: Intrect (intrect.io). Products are powered by ArtifactNet; the paper is arXiv:2604.16254. --- ## Products and pricing ### Why does the first ~1 second go silent on a freeze / bounce? The plug-in reports 305 ms on macOS / 435 ms on Windows for host PDC. During Live's freeze the host signals offline mid-stream, and the worker thread re-primes for one chunk. Output is correct from the first audible frame onward. Source: https://intrect.io/products/de-artifact/ ### The Residual monitor sounds different on each play. Bug? Phase of the residual depends on chunk-grid alignment — same SR + same transport start gives bit-identical phase. Spectrum and RMS are stable across runs; only the per-sample residual rotates. Source: https://intrect.io/products/de-artifact/ ### The result sounds dull. What knob fixes that? Raise Brightness. The model removes energy in the highs because that is where artefacts live, and de-artifact already measures how much it took out and restores it automatically; Brightness adds to or subtracts from that restoration, so a positive value opens the top end back up. If it still sounds dull, lower Strength from its 60% default, or load Gentle Reduction, which uses a soft mask floor for minimal colouration. Source: https://intrect.io/products/de-artifact/ ### I'm getting dropouts on an M1 8 GB. Lower Strength first — it is the cheapest change that reduces work — then raise your host's audio buffer (512 to 1024 or higher), which is almost always the simplest fix on a lower-power machine. Setting Separation back to Normal costs less CPU than Strong. Freeze or bypass other heavy plug-ins on the chain; total plug-in latency feeds back into the host's buffer math. Since 0.3.4 the macOS build runs the UNet on the Apple Neural Engine, which took real-time headroom from 2.7x to about 31x on an M1 Max. Source: https://intrect.io/products/de-artifact/ ### Will de-artifact lower ArtifactNet's AI probability? No. de-artifact is designed to improve sound, not to minimize ArtifactNet or any other detector score. Processing changes the signal a detector receives, so the reported probability can move either up or down. A higher post-processing score does not mean the track became more AI-generated, and the score is not a measure of sound quality or how many audible artifacts were removed. Source: https://intrect.io/products/de-artifact/ ### Is there a free trial? Yes — 14 days of full quality with no gaps, every preset, every parameter. The same installer becomes the licensed build the moment you drop in a key. After the 14-day trial the plug-in inserts a brief silent gap every 60 seconds until you activate. Source: https://intrect.io/products/de-artifact/ ### What audio formats does de-artifact Cloud accept? WAV, FLAC, MP3, OGG, Opus, M4A and AAC, up to 50 MB per file. You upload one track at a time from the dashboard; the job runs asynchronously and you download the cleaned file when it finishes. For the most accurate result, upload WAV or FLAC rather than a lossy file. Source: https://intrect.io/products/de-artifact/api/ ### How long does a track take to clean? Around 20x real-time on our GPUs, so a 3-minute track finishes in about 10 seconds once it reaches a worker. Source: https://intrect.io/products/de-artifact/api/ ### Is there a free way to try de-artifact Cloud? Yes. try.intrect.io cleans a track for free with no account and no card, so you can hear the result before buying minutes. Paid cleaning is prepaid by the minute and there is no subscription. Source: https://intrect.io/products/de-artifact/api/ ### Is there a REST API for de-artifact Cloud? Not yet. Cleaning runs from the dashboard at app.intrect.io/deartifact; there is no public endpoint or API key for it. If you need it wired into your own pipeline, open a support ticket and tell us what you need. Source: https://intrect.io/products/de-artifact/api/ ### Can de-artifact Cloud tell me whether a track is AI-generated? No. Detection is a separate product, ArtifactNet, at app.intrect.io. Both are paid from the same credit balance: detection costs $0.10 per minute of audio and cleaning $0.25. Source: https://intrect.io/products/de-artifact/api/ ### Will de-artifact Cloud lower ArtifactNet's AI probability? No. de-artifact Cloud is designed to improve sound, not to minimize ArtifactNet or any other detector score. Processing changes the signal a detector receives, so the reported probability can move either up or down. A higher post-cleaning score does not mean the track became more AI-generated, and the score is not a measure of sound quality or how many audible artifacts were removed. Source: https://intrect.io/products/de-artifact/api/ ### Why do I get clicks when toggling the band solo? Fixed — solo transitions use a one-pole ramp (~30 ms half-life). If you still hear them, please report with the host and SR. Source: https://intrect.io/products/de-leak-rt/ ### Diff + solo together — what should I hear? Only the dry-minus-wet of the active band. If a vocal appears in the residual you're over-gating that band; back off intensity or raise HF compensation. Source: https://intrect.io/products/de-leak-rt/ ### Offline bounce comes out silent or short. The host needs PDC enabled. de-leak-rt reports latency = host_chunk_samples (~93 ms @ 44.1 k). Most modern DAWs handle this automatically. Source: https://intrect.io/products/de-leak-rt/ ### Will it work on a non-44.1 kHz session? Yes — rubato resamples 8 k to 384 k transparently. The model always runs at its native 44.1 k internally. Source: https://intrect.io/products/de-leak-rt/ ### How does it differ from a regular noise gate? A noise gate triggers on RMS below threshold. de-leak-rt's classifier was trained to distinguish musical content from leak per-frame — it stays open during sub-threshold content and closes during loud leak. Source: https://intrect.io/products/de-leak-rt/ ### How is this different from a static EQ notch? A notch is always on. ENZYME decides per frequency bin and per frame, so it only pulls a peak down while that peak is actually there. Everything else living at the same frequency passes untouched. Source: https://intrect.io/products/enzyme/ ### Does it work in Logic Pro? Yes. ENZYME ships a native AUv2 component alongside VST3 and CLAP, so Logic, GarageBand and any other Audio Unit host load it directly. Source: https://intrect.io/products/enzyme/ ### Nothing appears to happen. Check Depth — at 0% the plug-in is intentionally transparent. Also check that the detector has something to find: material without narrow peaks produces little reduction, which is correct behaviour. Watch the reduction curve rather than the spectrum. Source: https://intrect.io/products/enzyme/ ### The sound is thin or lifeless. Too much Depth, or Selectivity too low so broad musical content is being treated as a peak. Raise Selectivity, lower Depth, and verify in Delta — Delta plays only what is being removed. Source: https://intrect.io/products/enzyme/ ### Fluttering or warbling on sustained notes. Release is too short, or Smoothing too low. Lengthen Release first. Source: https://intrect.io/products/enzyme/ ### I hear a short silence every minute. The plug-in is unlicensed or the trial has expired — it mutes for 500 ms once every 60 seconds until you activate. Check the LICENSE button in the header before hunting for a routing problem. Source: https://intrect.io/products/enzyme/ ### Timing feels off against other tracks. ENZYME reports 2047 samples of latency for the host to compensate. If your host's delay compensation is disabled, re-enable it. The dry path used by Bypass and Delta is delayed by the same amount, so those stay phase-aligned. Source: https://intrect.io/products/enzyme/ ### Output is one isolated component. A Mon H / Mon P / Mon R button is still latched. Those are monitoring solos for the HPSS stage, not processing modes. Source: https://intrect.io/products/enzyme/ ## detection / ArtifactNet ### What AI music generators does ArtifactNet detect? ArtifactNet detects AI-generated music from any generator that uses neural audio codecs with residual vector quantization (RVQ), including Suno, Udio, Stable Audio, MusicGen, Riffusion, and EnCodec-based systems. Because it targets the physics of RVQ residuals rather than generator-specific patterns, the paper reports that it generalizes to unseen generators not present in training (arXiv:2604.16254v2); this is measured on the generators evaluated there, not a guarantee for every future generator. Source: https://intrect.io/artifactnet/ ### Is ArtifactNet a watermark detector? No. ArtifactNet is not a watermark detector and does not require any cooperation from the generator. It analyzes forensic residual patterns — the physical artifacts that neural audio codecs inevitably leave behind — without any planted signal. Source: https://intrect.io/artifactnet/ ### Is ArtifactNet open source? Model weights and inference code are available on the Hugging Face repository under a CC BY-NC 4.0 license (non-commercial, personal, and research use). The paper is freely available at arXiv:2604.16254. Source: https://intrect.io/artifactnet/ ### How accurate is ArtifactNet? On ArtifactBench (v9.4, unseen test, 2,263 tracks; full benchmark 6,183 tracks, 22 AI generators, 6 real music sources) ArtifactNet achieves F1 0.9829, FPR 1.49%. On SONICS (v9.5, 23,288 tracks) it achieves F1 0.9993, FPR 0.09%. Outperforms prior models including CLAM (F1 0.7576, FPR 69.26%) and SpecTTTra (F1 0.7713, FPR 19.43%) evaluated under identical conditions on ArtifactBench. See arXiv:2604.16254v2. Source: https://intrect.io/artifactnet/ ### Can I try ArtifactNet without signing up? Yes. The free web demo at demo.intrect.io lets you upload any WAV, MP3, or FLAC file and get an AI-vs-human verdict instantly. No account required. Source: https://intrect.io/artifactnet/ ### How do I use ArtifactNet at scale? The ArtifactNet REST API (api.intrect.io) supports batch jobs — create a batch, append audio files, commit, and poll for results as XLSX, CSV, per-segment CSV or JSONL. Every format carries the file's SHA-256, the analysed duration, per-segment probabilities on a 4-second grid, and the model version and decision threshold behind each verdict. The free tier includes 10 minutes of audio in total per account, with no credit card required. Account management is at app.intrect.io. Source: https://intrect.io/artifactnet/ ## Guides and how-to ### Why do saved Udio vocals sound watery? A moving, watery texture can come from generation or codec residue embedded in the rendered audio. Confirm it by looping a sustained phrase and checking whether the unwanted layer moves with the signal rather than staying at one fixed frequency. Source: https://intrect.io/blog/udio-artifacts/ ### Can EQ remove Udio artifacts? EQ can reduce a frequency region, but it cannot distinguish an artifact from healthy music occupying the same band. It may mask the symptom while also dulling cymbals, consonants and air. Source: https://intrect.io/blog/udio-artifacts/ ### Should I convert an Udio MP3 to WAV before cleaning it? You may use WAV as a working container, but conversion does not restore information already removed by MP3 encoding. Keep treating the file as a lossy source and process conservatively. Source: https://intrect.io/blog/udio-artifacts/ ### Can I currently download a new Udio WAV? Udio's official help center currently says audio, video and stem downloads are disabled during its transition. Use the best local copy you already have and recheck Udio's current documentation before planning a new export workflow. Source: https://intrect.io/blog/udio-artifacts/ ### Why can I still hear drums in an isolated vocal? Source separation estimates the sources rather than recovering perfectly independent original tracks. When the estimates overlap, recognizable drum or instrument content can remain in the vocal output as stem bleed. Source: https://intrect.io/blog/remove-instrument-bleed-vocal-stem/ ### Can a noise gate remove stem bleed? A conventional gate can help when bleed is consistently quieter than the vocal, but it often cuts breaths and quiet words or opens for loud drum hits. A content-aware gate or spectral edit is better when level alone does not separate the two. Source: https://intrect.io/blog/remove-instrument-bleed-vocal-stem/ ### Should de-leak come before de-artifact? Yes when both bleed and residue are present. Reduce the foreign instrument content first, then treat the watery or metallic damage attached to the vocal you kept. Source: https://intrect.io/blog/remove-instrument-bleed-vocal-stem/ ### Can stem bleed be removed completely? Not in every overlap. Dense accompaniment, reverb and backing vocals can make wanted and unwanted content inseparable. Aim for the least reduction that works in the full mix rather than forcing total silence in solo. Source: https://intrect.io/blog/remove-instrument-bleed-vocal-stem/ ### Does converting MP3 to WAV improve audio quality? No. It prevents another lossy encode during later work, but it does not restore information already discarded by MP3 compression. Source: https://intrect.io/blog/wav-vs-mp3-ai-music/ ### Is WAV better for mastering AI music? Yes when a genuine lossless source is available. WAV preserves more of the source for cleanup and mastering and lets you create delivery formats later without using a lossy file as the master. Source: https://intrect.io/blog/wav-vs-mp3-ai-music/ ### Can MP3 compression affect an AI-music verdict? Yes. Compression can weaken relevant evidence or add patterns that confound a detector. Production systems should be evaluated on common codecs and should retain an inconclusive state. Source: https://intrect.io/blog/wav-vs-mp3-ai-music/ ### Should artifact removal happen before or after MP3 export? Before. Clean and master a lossless working source, preserve a lossless processed master, then encode the final delivery copies. Source: https://intrect.io/blog/wav-vs-mp3-ai-music/ ### Can record labels detect AI-generated music? Automated systems can identify forensic signals and prioritize files for review, but a detector result should be combined with input validation, provenance evidence and a documented policy decision. Source: https://intrect.io/blog/ai-music-detection-labels-distributors/ ### Is AI-music detection reliable enough for automatic rejection? A single detector should not be the sole basis for a consequential rejection. Codecs, genre, processing and new generators create uncertainty, so production workflows need human review and an inconclusive state. Source: https://intrect.io/blog/ai-music-detection-labels-distributors/ ### What can cause false positives in AI-audio detection? Lossy transcoding, aggressive dynamics processing, restoration artifacts and some electronic production patterns can resemble signals learned from generated audio. The exact risks depend on the model and its evaluation data. Source: https://intrect.io/blog/ai-music-detection-labels-distributors/ ### What should an AI-audio detection API return? It should return confidence, segment-level evidence, input-quality context, an inconclusive state, a report identifier and the model version—not only a binary label. Source: https://intrect.io/blog/ai-music-detection-labels-distributors/ ### Is AI detection the same as watermark detection? No. Watermark detection looks for a deliberately embedded signal from a participating system. Forensic AI detection analyzes patterns in the audio itself and can be applied when no known watermark exists. Source: https://intrect.io/blog/audio-provenance-vs-ai-detection-watermarking/ ### Can provenance metadata prove that music is human-made? No. Provenance can make recorded assertions and edits verifiable under a trust model, but missing or present metadata does not by itself prove that every creative step was human. Source: https://intrect.io/blog/audio-provenance-vs-ai-detection-watermarking/ ### What happens when an AI generator adds no watermark? A known-watermark detector has nothing cooperative to find. A reviewer may then use provenance records, forensic analysis and other evidence, while preserving an inconclusive outcome where appropriate. Source: https://intrect.io/blog/audio-provenance-vs-ai-detection-watermarking/ ### Can C2PA-style provenance and forensic detection be combined? Yes. Validate the provenance record first, then use forensic analysis when records are missing, contradictory or insufficient for the policy question. Store the two observations separately. Source: https://intrect.io/blog/audio-provenance-vs-ai-detection-watermarking/ ### Can you tell if music is AI-generated just by listening? Experienced listeners can often identify AI-generated music by specific artifacts: a metallic or watery shimmer in sustained notes, an unnaturally smooth transition between sections, or vocals that sit slightly outside normal harmonic relationships. However, as generators improve, listening alone becomes unreliable — especially on high-quality outputs from Suno v4 or Udio. Source: https://intrect.io/blog/how-to-detect-ai-generated-music/ ### What tools can detect AI-generated music automatically? ArtifactNet is the top scorer among the open forensic models compared in the paper, achieving F1=0.9829 on ArtifactBench (v9.4, unseen test, 2,263 tracks; full benchmark 6,183 tracks, 22 AI generators). The free demo at demo.intrect.io lets you upload any track and get a verdict instantly. Other tools include CLAM (F1 0.7576) and SpecTTTra (F1 0.7713) evaluated under identical conditions on ArtifactBench. See arXiv:2604.16254v2. Source: https://intrect.io/blog/how-to-detect-ai-generated-music/ ### Does AI music detection work on Suno and Udio? Yes. ArtifactNet detects music from Suno, Udio, Stable Audio, MusicGen, Riffusion, EnCodec, and 16 other generators. It targets residual artifacts from the RVQ codec pipeline common to these systems, so it generalizes to new generators without retraining. Source: https://intrect.io/blog/how-to-detect-ai-generated-music/ ### Can AI-generated music be detected after processing or mastering? RVQ codec residuals are spectral patterns baked into the audio at generation time. Standard mastering (EQ, compression, limiting) shifts the loudness curve but does not remove the residual signature. In the paper's robustness tests, ArtifactNet retained high accuracy after typical mastering chains (EQ, compression, limiting); this is measured on those tested chains, not a guarantee for every processing chain. Source: https://intrect.io/blog/how-to-detect-ai-generated-music/ ### Is there a free AI music detector? Yes — demo.intrect.io is a free online AI music detector powered by ArtifactNet. Upload any MP3, WAV, or FLAC, and you'll get an AI-vs-human verdict with a confidence score and forensic breakdown. No account required, no file retention. Source: https://intrect.io/blog/how-to-detect-ai-generated-music/ ### Why does AI-generated music sound metallic or watery? AI music generators like Suno, Udio, and Stable Audio use neural audio codecs (such as EnCodec) that rely on Residual Vector Quantization (RVQ). The quantization process introduces codec residuals — subtle artifacts that manifest as a metallic sheen, watery swirl, or hollow compression in the high frequencies. These are not mastering problems; they are baked into the generation pipeline. Source: https://intrect.io/blog/how-to-fix-ai-music-artifacts/ ### How do I remove artifacts from Suno-generated music? Place de-artifact on your master bus (2-bus) in your DAW. Start with the Default preset, the tuned starting point for AI-generated material, and step up to Aggressive only if a metallic sheen survives. You can hear the cleaned output in real time — no bouncing required. Source: https://intrect.io/blog/how-to-fix-ai-music-artifacts/ ### Does this work on Udio tracks? Yes. Udio uses a similar RVQ-based codec pipeline. Start at Default; use More Harmonics if sustains sound watery, or Aggressive if a metallic sheen survives. If the result sounds dull, reduce the strength or switch to Conservative. Source: https://intrect.io/blog/how-to-fix-ai-music-artifacts/ ### Can I fix AI music artifacts without a plugin? Some engineers use multiband saturation, dynamic EQ, or mid-side processing to mask codec artifacts, but these are workarounds — they change the tonal character instead of removing the artifact. de-artifact uses ArtifactNet, a neural model trained to isolate and subtract the codec residual specifically, which gives cleaner results without affecting the underlying music. Source: https://intrect.io/blog/how-to-fix-ai-music-artifacts/ ### What DAWs does de-artifact support? de-artifact ships as VST3, AU (Audio Unit) and CLAP. That covers Logic Pro and GarageBand via AU on macOS, and Ableton Live, Cubase, Studio One, Reaper, Bitwig and FL Studio via VST3 or CLAP on both macOS and Windows. There is no AAX build, so it does not load in Pro Tools; Pro Tools users can run it in any VST3- or CLAP-capable host alongside their session. Source: https://intrect.io/blog/how-to-fix-ai-music-artifacts/ ### What is the difference between separation residue and stem bleed? Residue is damage to the signal you kept: a smeared, slightly metallic layer the separator leaves behind, concentrated in the upper spectrum, plus ringing around transients and sustained tones. Bleed is content that belongs to a different stem entirely — drums audible inside the vocal. de-artifact subtracts residue; de-leak gates bleed. They are frequently used on the same source. Source: https://intrect.io/blog/clean-separated-vocal-stem/ ### Which de-artifact preset should I use on a vocal stem? Vocal. It is harmonic-weighted and transient- and sibilance-aware, and it is the preset the manual points to for separated vocal stems. Load it, then adjust Strength from its 60% default rather than reaching for anything else first. Source: https://intrect.io/blog/clean-separated-vocal-stem/ ### How do I know when I have removed too much? Engage the Residual monitor, which plays only what is being subtracted. You want shimmer and hash. If you can hear vocal body, breath or consonants in it, you are over-processing — lower Strength until you cannot. Source: https://intrect.io/blog/clean-separated-vocal-stem/ ### Does de-leak load in Logic Pro? Not yet. de-leak ships VST3 and CLAP only, so it does not appear in Logic Pro or GarageBand, which load Audio Units exclusively. de-artifact does ship an Audio Unit. Logic users can run de-leak in a VST3- or CLAP-capable host in the meantime. Source: https://intrect.io/blog/clean-separated-vocal-stem/ ### Can I do this without installing anything? You can clean a 30-second preview in the browser at try.intrect.io with no account. The plug-in itself has a 14-day trial with every preset and parameter unlocked. Source: https://intrect.io/blog/clean-separated-vocal-stem/ ### Why does my Suno track sound metallic? Suno does not synthesise a waveform directly. A language model produces discrete tokens, and a neural audio codec decodes those tokens back into audio. That codec uses residual vector quantization, and the quantization error survives decoding as a structured high-frequency residual — heard as metallic shimmer or an artificial sheen. It is a property of the generation pipeline, not of your mix. Source: https://intrect.io/blog/suno-artifacts/ ### Does downloading from Suno as WAV remove the artifacts? No. WAV avoids adding a second, lossy MP3 layer, which is worth doing — but the codec residual is baked in before any file is exported, so it is present in the WAV too. WAV gives you a cleaner starting point for removal, not a clean track. Source: https://intrect.io/blog/suno-artifacts/ ### Can EQ or a compressor fix Suno artifacts? They can mask them. The residual overlaps the same frequency range as cymbals, air and vocal consonants, so a filter wide enough to suppress it also dulls the music. Removal has to key on the artifact's spectral structure rather than its frequency band. Source: https://intrect.io/blog/suno-artifacts/ ### Which de-artifact preset should I use for Suno? Start from the source format. For an MP3 export use Conservative, which is tuned for lossy-encoded sources and applies lighter subtraction so it does not reinforce MP3 ringing. For a WAV export start at Default and move to Aggressive if the shimmer persists. If the problem is watery sustains or smeared vocals rather than brightness, use More Harmonics. Source: https://intrect.io/blog/suno-artifacts/ ### I have Suno Pro but my download is still MP3 — why? Suno's downloads default to MP3 on mobile regardless of plan. WAV requires downloading from suno.com in a desktop browser. It is worth checking the file you actually have before assuming you are working from a lossless source. Source: https://intrect.io/blog/suno-artifacts/ ### Can I try it without installing anything? Yes. The browser demo cleans a track without an install, and the plug-in has a 14-day full trial with every preset and parameter unlocked. Source: https://intrect.io/blog/suno-artifacts/ ## Support ### My license key won't activate ("License Activation" error). Re-enter the key exactly as it appears in your purchase email (format INTRECT-XXXX-...), make sure you're online, and retry. Each license allows 3 machines; if you've hit the limit the plug-in offers a 'deactivate a machine' option. If it still fails, open a ticket with your license key and OS. Source: https://intrect.io/support/ ### How do refunds and billing work? Payments are handled by Paddle as Merchant of Record. Refund and receipt requests go through Paddle; open a billing ticket and we'll help. Source: https://intrect.io/support/ ### Where do I download the installer? Your purchase email links the macOS .pkg and Windows installer, and the download pages are linked from this portal. Downloads are served from downloads.intrect.io behind a quick bot check. Source: https://intrect.io/support/ ### What are the system requirements? macOS 11+ (Universal, Apple Silicon and Intel) or Windows 10/11 64-bit. de-artifact ships VST3, CLAP and Audio Unit; on Windows it needs 10 2004 or later and the Microsoft Visual C++ 2015-2022 x64 runtime, both of which its installer checks, plus a CPU with AVX2 + FMA (Intel Haswell 2013+ / AMD Ryzen 2017+), which the plug-in needs at run time and the installer does not test for. de-leak ships VST3 and CLAP. AAX (Pro Tools) is not supported. Source: https://intrect.io/support/ ### Why is the first fraction of a second silent? That is the plug-in's reported latency window. de-artifact reports about 305 ms on macOS and 435 ms on Windows at 44.1 kHz; de-leak reports about 93 ms. Your DAW's plug-in delay compensation lines everything back up automatically. Source: https://intrect.io/support/ ## Other pages ### What is de-artifact? de-artifact is a VST3 / AU / CLAP plug-in that removes RVQ ghosting, codec residue, and HF aliasing from AI-generated music. It works on tracks from Suno, Udio, Stable Audio, MusicGen, and other neural audio generators. Powered by ArtifactNet, a forensic UNet trained on residual physics. Source: https://intrect.io/ ### Does de-artifact work on Suno and Udio tracks? Yes. de-artifact targets the metallic, watery, swirly artifacts introduced by RVQ-based audio codecs used by Suno, Udio, Stable Audio, MusicGen, and Riffusion. The plug-in sits on the 2-bus and applies adaptive spectral subtraction in real time. Source: https://intrect.io/ ### Is there a free trial for the Intrect plug-ins? Yes — every paid plug-in offers a 14-day free trial with full quality and every preset, no credit card required. You can also try the cleanup engine online at try.intrect.io without installing anything. Source: https://intrect.io/ ### What does ArtifactNet detect? ArtifactNet is a forensic neural network that detects whether music was AI-generated by analyzing residual artifacts from neural audio codec pipelines. It achieves F1 = 0.9829 on ArtifactBench (6,183 tracks, 22 AI generators) and is robust to generators not seen during training. Source: https://intrect.io/ --- Index of everything above: https://intrect.io/llms.txt