Terms of Service

Rules for using Lyrictomelody AI

Software-as-a-Service Agreement

Use Lyrictomelody AI responsibly

These terms describe how you may access and use Lyrictomelody AI's tools -- including lyric-to-melody generation, AI singing and voice model training, voice conversion, vocal separation, multi-track and symbolic-music generation, and MIDI-to-vocals -- how credits are billed, and how your inputs and outputs are handled.

Last updated: October 12, 2026

Purpose

AI-assisted songwriting: melody, vocal, voice-model, and instrumental generation tools for music creators.

Eligibility

You must be 18+ and use the service for legitimate business or personal workloads.

Billing

Credits are prepaid (subscription or one-time) and deducted when a job is accepted; failed jobs are refunded automatically. Cost depends on the tool and, for some tools, on audio length.

Your responsibilities

Upload or input only content you have permission to process. You confirm this each time you submit an upload-based job, and we keep a record of that confirmation.
Voice training and voice conversion: only submit a voice recording that is your own, or one you have explicit permission from that person to use for training or conversion. You confirm this at the point of training and again at each voice conversion (including the Voice Reference clip). See "Voice, likeness & digital replicas" below.
Uploaded songs and recordings: only process music you own or are licensed to use. See "Copyright, uploaded music & takedowns" below.
Do not upload malicious content, PII you lack consent for, or regulated data that requires special handling (HIPAA, PCI).
Respect rate limits and do not attempt to reverse-engineer or resell the service.
Keep your account credentials secure and notify us of suspicious activity.

Acceptable Use Policy — prohibited content

Sexual or pornographic content (NSFW), including sexually explicit lyrics, vocals, or audio.
Violent or graphic content that glorifies, incites, or depicts gore, self-harm, or real-world harm.
Hate speech or content that harasses, dehumanizes, or promotes discrimination or violence against protected groups.
Child sexual abuse material (CSAM) or any content that sexualizes minors — zero tolerance, reported to authorities.
Deepfakes and impersonation — cloning a real person's voice or likeness, or generating songs attributed to real artists, without their explicit consent.
Content that infringes copyright, trademarks, or other intellectual property, including reproducing protected lyrics, melodies, or brands you do not own or have rights to.

Enforcement, reporting & content moderation

Enforcement: violations may result in content removal, credit forfeiture without refund, feature restrictions, and account suspension or permanent termination. Illegal content (e.g. CSAM) is preserved as required and reported to law enforcement.
Reporting: report abusive, infringing, or policy-violating content to support@lyrictomelodyai.com or through our contact page. Copyright holders should follow the procedure on our Copyright & DMCA page (/copyright); individuals whose voice or likeness was used without permission should follow the procedure in "Voice, likeness & digital replicas" below.
Moderation: we combine automated safety filters on inputs and generated outputs with human review of reports. We may review, restrict, or remove content and reserve the right to refuse generations that violate this policy.

Credits and payments

All paid plans are processed through our payment processor.
Credits do not expire but are non-transferable between accounts.
Credit cost depends on the tool and, for some tools, on the length of your audio or melody (for example, Voice Reference conversion and MIDI to Vocals use length tiers). The exact cost is shown before you start processing.
If a job fails on our side or at a provider, the credits for that job are refunded automatically.
Subscriptions renew automatically until canceled. You can cancel from your dashboard at any time; cancellation stops future renewals and your plan stays active until the end of the paid period. If you cannot cancel yourself, email support and we will cancel it for you.
If you are a consumer in the EU/UK, you may have a statutory right to withdraw within 14 days; where we start delivering digital content or credits immediately at your request, you acknowledge that this right may be lost once delivery has begun, to the extent permitted by law. See the Refund Policy.
Charge disputes should be sent to support before filing with your card issuer (see Refund Policy).

Data handling

We store submitted lyrics and generated outputs only as needed to provide downloads during the retention period.
Operational logs include metadata (timestamp, runtime, settings, input length) but do not include full lyric content.
See the Privacy Policy for full details, including your right to delete account data at any time.

Termination

We may suspend accounts that abuse the platform, attempt fraud, or violate export/control laws.
You may cancel anytime; unused credits can be refunded per the Refund Policy (where applicable).
Deleting your account disables access immediately. On your request (support@lyrictomelodyai.com) we then permanently erase your files, voice models and profile data within 30 days, except records we must keep for legal or dispute purposes (for example invoices and rights-confirmation records).

AI models & technology disclosure

Song and instrumental generation uses third-party models including MiniMax Music, Google Lyria 3 Pro, Meta MusicGen, ACE-Step and ElevenLabs Music, delivered through Replicate and fal.
Song With My Voice writes a complete song from your lyrics and style choices with MiniMax Music (delivered through Replicate), then converts the singing so it follows the recording you upload, using SoulX-Singer-SVC (singing voice conversion) that we run on Modal, and mixes the result with the instrumental. The converted voice is based on your recording and may not sound identical in every passage; quality depends on the recording's clarity and range. We do not keep a reusable copy of your voice.
Text to Singing Voice for lyrics in Spanish, Japanese, Korean, Portuguese, German, Italian, French or Russian writes a full song from your lyrics with Mureka 9.5 (delivered through fal) and then isolates the vocal with Demucs (via fal). The vocal is sung by an AI voice; your own voice recording is not used in this mode. Because the vocal is separated from a full song, faint backing may remain and the sung words may differ from your lyrics. English lyrics keep using the SoulX-Singer flow with your own voice sample.
Finish This Melody turns your selected melody and lyrics into a complete song with Mureka 9.5, delivered through fal. The song follows your melody as a musical guide; exact notes, timing and phrasing may vary.
Melody to Song Generator turns the MIDI melody you upload and your lyrics into a complete song with Mureka 9.5, delivered through fal. The song follows your melody as a musical guide; exact notes, timing and phrasing may vary.
Reference Audio to Song writes a new song from your lyrics, guided by a 30-second clip of audio you choose, with Mureka 9.5 (delivered through fal). The result is a new interpretation, not a copy of your recording: its melody, arrangement and vocals may differ. Only submit audio you own or have permission to use.
Lyrics to Backing Track writes a full song from your lyrics with Mureka 9.5 (delivered through fal), then separates it with Demucs (via fal) into an instrumental, guide vocals and the full song. Separation is imperfect: faint traces of the vocal can remain in the instrumental.
AI Vocal Hooks for Beatmakers places an AI singing vocal on the beat you upload. Your beat is stored with your other uploads and used only to build the demo mix and the aligned vocal file you download; it is not changed or used to train models, and it is not sent to the AI vocal service. Only upload a beat you own or have permission to use, and check your rights in the beat, lyrics, melody and voice before releasing the result. The vocal can differ from the notes and timing of your melody.
AI DJ Drop turns a name and a few words you enter into a short audio clip with a spoken voice, music and sound effects, using Seed Audio 1.0 (a ByteDance model, delivered through fal). A name can be pronounced differently from how you say it, the length is not guaranteed, and the result is an AI-generated draft you should listen to before you use it. It uses built-in voices only: you cannot upload a voice or have a voice cloned. Only enter names and words you have the right to use, and check any trademark or advertising rules that apply to where you publish the clip.
AI Instrumental Generator creates instrumental music from a text description with Mureka 9.5 (delivered through fal). The result can differ from the style, instruments and length you describe.
Music Extender and Extend Song use Stable Audio 3 from Stability AI, delivered through fal. Powered by Stability AI. This Stability AI Model is licensed under the Stability AI Community License, Copyright © Stability AI Ltd. All Rights Reserved.
MIDI-to-vocals, lyric-rewrite (Change Lyrics) and melody transcription use SoulX-Singer models that we run on Modal. Voice Reference conversion uses SoulX-Singer-SVC (singing voice conversion), also on Modal.
Trained and preset voice conversion uses RVC v2 (Retrieval-based Voice Conversion), an open-source model, via Replicate.
Vocal remover and stem separation use Demucs; lyric alignment and transcription use ElevenLabs forced alignment and Whisper via fal.
Melody transcription uses the NVIDIA Parakeet-TDT 0.6B v2 speech model (licensed under CC BY 4.0, © NVIDIA) inside the SoulX-Singer preprocessing pipeline.
AI lyric and text generation uses OpenAI GPT models.
Underlying models, versions and providers may change as we improve quality; this page is updated to reflect the models in use. Your inputs are sent to these providers only to process your request (see the Privacy Policy for the list of processors).
AI-generated and AI-converted audio is synthetic. You are responsible for disclosing that it is AI-generated or AI-converted wherever disclosure is required by applicable law, distributor or platform rules (for example, EU AI Act transparency obligations for synthetic audio, or streaming-platform AI policies).

Voice, likeness & digital replicas

Voice features: Trained Voice (a model trained from a recording you submit, or a preset voice) and Voice Reference (a short clip of a voice used as a one-off reference for a conversion). A person's voice can identify them; treat every voice input as personal data of that person.
Permission: you may use only (a) your own voice, or (b) the voice of someone who has given you explicit, informed permission for this specific use. If you intend to publish, distribute or sell the result, that permission must expressly cover commercial use.
Prohibited: creating or distributing output that imitates a real, identifiable person (including public figures, recording artists, and deceased persons whose estate controls their voice or likeness) without their permission; any use that is deceptive, fraudulent, defamatory, sexual, political, or that implies endorsement, authorship or a statement the person did not make.
Confirmation record: before each voice training or voice conversion you must confirm the above. We record your confirmation (account, time, wording version and which inputs it covered) and may use the record to investigate complaints.
Complaints and removal: if your voice or likeness was used without your permission, email support@lyrictomelodyai.com with links or job details and a way to verify your identity (or that you represent the person or their estate). We will investigate promptly, and may remove the voice model, the reference clip and the outputs, notify the account holder, and suspend or terminate the account responsible. False reports may be refused.
We do not provide a way to publish a trained voice for other users to select, and we do not offer preset voices that imitate real people.

Copyright, uploaded music & takedowns

Your content: songs, recordings, stems, lyrics and MIDI you upload remain yours or your licensors'. You grant us a limited, non-exclusive license to host and process them solely to provide the tool you requested.
Clearance: only upload music you own or are authorized to use for the specific processing requested (for example stem separation, voice conversion, lyric rewriting or extension). Running a commercially released recording that you do not own or license through these tools, and then publishing or distributing the result, may infringe copyright and related rights of the recording, composition and performer. You are solely responsible for obtaining clearances.
Confirmation record: each upload-based tool asks you to confirm that you have the necessary rights before processing. We record that confirmation.
Outputs: as between you and us, you may use outputs subject to these Terms and to any third-party rights in the inputs you provided. Outputs are generated automatically; they may not be protected by copyright in some countries, may resemble existing works, and we give no assurance of originality or non-infringement.
Takedowns: copyright holders can report infringement, send a DMCA notice or counter-notice, and read our repeat-infringer policy on the Copyright & DMCA page (/copyright). We remove or disable access to reported material, and terminate accounts of repeat infringers in appropriate circumstances.

Trained voice models

Voice models you train are private to your account. We do not currently offer a way to publish a trained voice model for other users to select and use.
To request deletion of a trained voice model, contact support@lyrictomelodyai.com; this does not retroactively affect outputs already generated with it.

Disclaimer of warranties

The service, including all AI-generated audio, lyrics, and outputs, is provided "as is" and "as available" without warranties of any kind, whether express or implied.
We disclaim all implied warranties, including merchantability, fitness for a particular purpose, non-infringement, and any warranty that outputs will be accurate, original, uninterrupted, or error-free.
AI outputs are generated automatically and may be inaccurate, unexpected, or resemble existing works. You are responsible for reviewing outputs and confirming you have the rights to use them before publishing or distributing.

Limitation of liability

To the maximum extent permitted by law, Lyrictomelody AI and its suppliers are not liable for any indirect, incidental, special, consequential, or punitive damages, or any loss of profits, data, goodwill, or business, arising from your use of the service.
Our total aggregate liability for any claim relating to the service is limited to the amount you paid to us in the twelve (12) months preceding the event giving rise to the claim.
Some jurisdictions do not allow certain limitations, so parts of the above may not apply to you; in that case liability is limited to the smallest amount permitted by law.

Governing law & contact

These terms are governed by the laws of the People’s Republic of China. Disputes are handled in Shanghai jurisdiction. Nothing in these terms limits any non-waivable consumer protection, refund, or cancellation right available to you under the mandatory law of your country of residence, where such rights cannot legally be waived by agreement.

Questions? Email support@lyrictomelodyai.com or visit our contact page.