Brand Jingle & Theme Studio

Problem

When a brand needs an ad jingle, store loop, podcast intro, or sonic logo, generating a full track immediately is risky: style drifts, duration is hard to control, spoken slogans get mistaken for sung lyrics, and paid generation calls are retried blindly. This skill turns the workflow into reviewable sketches, frozen payloads, and derived durations.

Workflow and boundaries

• Inputs: require a brand or product name, placement, language, and route: sung hook, instrumental, or spoken slogan. Reuse known slogans, required words, mood, genre, vocal direction, and exclusions.
• Sketch first: produce three style sketches on one primary duration by default; about 15s for ads and podcast intros, 30s for radio and store loops, and 2–5s for sonic logos. Derive other durations only after the user locks one.
• Cost points: copy, proposals, and style cards are free; each beatra.music.generate is billed; spoken slogans use a separate billed beatra.speech.synthesize. Confirm prompt, lyrics/instrumental flag, track name, model, max cost, and client_request_id before submission.
• Parameter safety: use model: "suno-5.5" for standard music and do not silently switch to auto; there is no duration parameter, so target duration is written into the prompt; include model_options.suno.custom_mode: true when the caller supplies lyrics; set instrumental: true for sonic logos.
• Deliverable split: instrumental bed and spoken slogan must be separate deliverables; accepted sketches are not mutated; derived durations are new billed work.
• Verification: poll with beatra.tasks.get and report real duration_seconds, MIME, size, URL or artifact ID, resolved model, and billing.net_charged_credits; suggest concrete next steps such as adjusting BPM or tightening the ending.
• Boundaries: use adjacent skills for personalized songs, lyrics-to-song, short-video BGM, livestream packs, or covers; treat requested duration as an orchestration target, not a sample-exact guarantee.