demorphus
If a distributor rejects your AI music, save the rejection message and identify what it asks you to fix. Audio quality, AI classification, rights and release metadata are different issues. Process the track only when the problem calls for an audio change, then review the exact file you plan to resubmit.
Read the rejection before choosing a fix
Record the release identifier, affected track, date, distributor and exact wording. A message about a file upload is not the same as a content-policy decision. If the reason is unclear, ask support which requirement the release did not meet and what evidence or correction it needs.
Do not infer the distributor’s detection engine or threshold from an unexplained rejection. A SH Labs report gives you an external measurement; it does not reveal the systems used by every distributor.
Match the action to the problem
| Reason stated | Useful next step |
|---|---|
| Audio format or damaged file | Check the source, export settings and upload requirements. Listen to the full exported file. |
| Audible distortion or poor sound | Locate the defect and compare a targeted repair with the original. |
| AI classification or provenance | Read the service’s AI policy, retain production records and analyze the exact submitted audio. |
| Rights, identity or impersonation | Provide the requested permissions or correct the release. A new audio score cannot supply them. |
| Credits, title or artwork | Correct the relevant release fields rather than changing the sound. |
When Cleaner is relevant
If your objective is to change the audio cues recognized by a detector, start with the Suno fingerprint workflow. Keep a baseline report, process your source and compare the returned SH Labs verdict with the original. The Studio prepares the processed file and its report for download.
Demorphus shows the measured classification, including when the desired result is not reached. A Human Made response can be useful for evaluating the processed file; it is not a universal distribution clearance. Use the internal results and method to understand the evidence before choosing a plan.
When another treatment is the wrong next step
Do not repeatedly process a file when the rejection is about permissions, identity, artwork or credits. That changes the audio while leaving the stated issue unresolved. It also makes it harder to track which version was submitted and which one received a score.
For an audible fault, work on the relevant section or source where possible. The artifact guide separates vocal glitches, noise and transients. A lower AI score alone is not evidence that a damaged vocal or clipped chorus has been repaired.
Keep a record when you resubmit or appeal
Use one folder per release with the source, selected export, report and distributor correspondence. Note what changed and why. If support asks for production evidence, provide the appropriate records rather than presenting a detector label as proof of human authorship.
Check whether the service asks you to edit the existing submission or open a support case. Follow its instructions instead of sending multiple conflicting versions. Keep the final delivery confirmation: it measures a different outcome from the detector report.
Does DistroKid reject all AI music?
No. DistroKid’s public AI guidance allows AI-created music under conditions and notes that streaming services can reject releases that do not meet their rules. Read the policy and your specific rejection together. Another distributor may apply different requirements.
For your next release, use the Suno distribution workflow before uploading. It connects source selection, processing, analysis and submission records without treating them as the same check.
Policy sources checked October 30, 2026. Check the current rules for your chosen distributor before submitting.
Prepared with AI assistance. Sources below.