Use Sonilo’s single catalog model for focused music briefs such as narration beds and light soundtrack sketches. WaveSpeed exposes a duration parameter, but the owner-audited EZ Ai Assist interface does not: it forwards a fixed 60-second request rather than offering a track-length selector.
Editorial starting points, not a performance ranking. Check current app access before choosing a model.
Leave space for a spoken explanation
WaveSpeed describes background scoring from a natural-language style and arrangement brief. For narration, begin with a restrained instrument palette, gentle dynamics and a small motif that does not compete with the speaker.
Tradeoff: No dedicated instrumental-only switch is established for the audited app. Wording such as no vocals is a creative request, so listen for singing or speech before using the result.
Use the same model for a contrasting musical direction by changing instrumentation, energy and arrangement. The host describes soundtrack ideation and content scoring, making a tightly scoped mood sketch a reasonable evaluation task.
Tradeoff: This is not a separate Sonilo variant or a claim of guaranteed speed. Avoid copying a named artist, and assess whether the returned theme fits the project rather than assuming a prompt makes it unique or cleared.
Official documentation describes the named model or gateway—not every feature available in this app.
Prompt and duration are explicit inputs
WaveSpeed’s sonilo/text-to-music schema requires prompt and integer duration. It documents 1–360 seconds with a default of 30. That is the gateway’s target-duration range, not a user-selectable range in EZ Ai Assist; the app audit reports an explicit duration of 60 instead of the host default.
The reviewed Sonilo schema describes style, mood, instruments and arrangement in the prompt. It does not establish dedicated lyrics, instrumental, stem-export, seed or audio-upload parameters for this route. Describing separate instruments or asking for a loop does not create separate tracks or a validated loop boundary.
WaveSpeed describes asynchronous submission, a prediction ID and result retrieval. The model page’s timing estimates are not service guarantees, and the documented duration value is a request parameter rather than an independent measurement of the output. Inspect the final audio file before planning cuts, fades or narration timing.
Every existing catalog entry for this provider is linked below. Status labels come from the individual website guides; catalog coverage is not a guarantee of current app access. Historical and snapshot pages remain references, not extra active models.
Original prompt examples, not generated-output demonstrations. They have not been generation-tested.
Workflow 1
Score an instructional voiceover
Write a concise brief within the app’s reported 500-character prompt allowance. Set the musical role first: support a speaker, not draw attention away from the explanation. There is no duration selector to adjust.
Compose an original understated instrumental for a practical home-repair tutorial. Use muted guitar, a soft rounded bass and light brush percussion. Keep the pulse steady, leave room for speech and develop one small friendly motif. Avoid vocals, sudden accents, a dramatic build or a long introduction. Finish with a gentle cadence.
Check the result: Listen beneath a sample of narration at the intended level. Check for unwanted voices, distracting peaks and an awkward ending; measure actual duration before editing.
Change the musical palette rather than adding many competing styles. Keep a note of the exact brief and what worked, so the next iteration has a clear purpose instead of being an unstructured retry.
Create an original gentle electronic cue for a neighborhood science-project update. Use soft marimba-like notes, airy synth chords and a restrained bass pulse. Make the mood curious and welcoming, with space between phrases and controlled dynamics. Avoid singing, speech, heavy drums and recognizable melodies; end softly rather than stopping mid-phrase.
Check the result: Evaluate whether the mood matches the message and whether the cue stays unobtrusive. Do not assume a soft ending is a seamless loop or that instruments can be downloaded as stems.
WaveSpeed’s Sonilo documentation was checked independently. App controls and the forwarded 60-second value come from the owner-supplied October 9, 2026 music audit, not inspected application handlers or a paid generation test. A documented request limit does not guarantee actual returned length.
Owner audit · October 9
Controls visible in the audited app
The app uses the main music prompt with a 1–500-character allowance. No separate lyrics field, instrumental toggle or duration selector is reported for Sonilo. The presence of these ideas in a music brief is not proof of a corresponding parameter.
Owner audit · October 9
Values sent upstream
The audit records prompt and duration = 60 sent to WaveSpeed’s sonilo/text-to-music. Its wrapper clamp is 10–180 seconds, but users do not select that range. The fixed minute is a request, not a guarantee of exact output length.
Audit vs. current host schema
How the earlier duration label fits
The earlier three-minute picker description is narrower than the current 360-second gateway ceiling. Neither is an app duration control. This guide documents the difference without changing the picker, wrapper, provider request or billing behavior.
Before you rely on a result
Limits, costs and tradeoffs
The reviewed route does not establish a guaranteed instrumental-only mode or exact lyric reproduction. If the result contains unwanted voices or recognizable material, do not publish it merely because the prompt asked otherwise.
A requested minute may need trimming or fades for a specific edit. Seamless looping, exact beats per minute and individual instrument stems were not independently verified.
The gateway documents duration-based charging. An old flat per-track picker estimate, provider hosting cost and actual EZ Ai Assist customer charge are different things; this page does not quote an app price.
Compare another prompt-led music option, with its AIML route and possible MiniMax fallback explicitly disclosed.
Common questions
Sonilo, without the assumptions
Can I select a six-minute Sonilo track?
No duration selector is reported in the app. Six minutes is the documented WaveSpeed maximum; the audited integration forwards 60 seconds. Provider limits and the wrapper’s internal clamp are not user controls.
Is the output guaranteed to last exactly 60 seconds?
No. Sixty seconds is the owner-audited upstream request value. No generated file was measured during this review, so actual timing remains an output check before editing or publishing.
Why are there two starting points but one model?
They are two practical uses of Sonilo Text-to-Music: narration support and mood exploration. The website keeps one model entry and does not invent a second version to fill the comparison section.
Highlighted provider contracts reviewed . Individual model guides carry their own review dates. Sources can change; these pages do not perform live health checks.
WaveSpeed’s Sonilo documentation was checked independently. App controls and the forwarded 60-second value come from the owner-supplied October 9, 2026 music audit, not inspected application handlers or a paid generation test. A documented request limit does not guarantee actual returned length.