ElevenHacks · Shipped · Live

Stan Bran

The New Product Music Man.

“Far out new product, man, tell me more about it.”

40 custom-branded audio assets. One small business. Five minutes. Stan Bran is the jingle guy you would’ve hired in 1982 if you were opening a laundromat — home studio packed with synths, rates you could actually afford, radio spot by Friday. Now he lives on the ElevenLabs stack and he works for every small business on the internet.

40Assets Per Kit
71Voices
5 minTurnaround
$4.99Or Free
Visit stanbran.com See how he works →
Meet the man

He’s been doing this since 1982.

You’ve met Stan before. He’s the guy in the Hawaiian shirt who ran the local jingle shop out of his garage. Made the radio spot for the car dealership, the hold music for the dentist, the voiceover for the TV ad. Rates you could actually afford. We just gave him the ElevenLabs stack.

stan — home-studio
Stan Bran at his home studio, surrounded by vintage synths and keyboards
Home studio · vintage synths · half a century of jingle work “Tell me more about it.”
Three things at the core

One form. Forty assets. Five minutes.

01

He writes.

Claude Sonnet 4 takes your product brief — name, tagline, description, business type, tone — and writes 26 voiceover scripts, 4 music prompts (genre, tempo, instrumentation, structure), and 10 SFX prompts. All in the brand’s voice. All in one model call. No templates.

02

He performs.

ElevenLabs generates everything in parallel — TTS across 71 voices with per-preset stability and style, instrumental music beds via the Music API, designed SFX via the Sound Effects API. Two at a time, streamed as MP3 buffers, cached in Postgres, ready to ship.

03

He ships.

Signed download URLs, 30-day expiry, HMAC-verified. Welcome email through Resend. Stripe receipt. Preview page with a built-in Web Audio mixer for layering music + VO + SFX before you even leave the tab. Export as WAV. Ship the ad.

What’s in the box

40 assets. Every corner of the launch.

Stan doesn’t do a “voiceover.” He does the whole kit — the stuff a $5,000 studio session would hand you after three weeks. Tailored to your tone, your industry, your audience, your features. Full commercial rights. Yours forever.

Voiceovers

26

Teasers (hype / mystery / countdown), explainers (60s / 30s / deep-dive), commercials (15s / 30s / 60s), radio ads, five flavors of testimonial, podcast pre-roll and mid-roll, social hooks, on-hold messages, IVR greetings, event intros.

Music Beds

4

Brand anthem, intro bed, commercial bed, upbeat social short. All instrumental, all generated via the ElevenLabs Music API with detailed prompts — genre, BPM range, instrumentation, production style, key feeling, reference genre.

Sound Effects

10

Logo reveal stingers, notification chimes, success celebrations, transition whooshes, risers, brand ambient textures, applause. Prompt influence tuned per category so a logo reveal doesn’t sound like a notification.

Voice Roster

71

Full ElevenLabs voice library accessible, with 8 curated presets (female professional, female energetic, female warm, female confident, male deep, male friendly, male narrator, male energetic) for quick picking. Or bring your own.

Business Types

60+

Coffee shops to chiropractors, plumbers to pizza places, law firms to nail salons. Industry-aware generation means a yoga studio sounds like a yoga studio, not like an auto shop. Library covers 475 verticals total.

Commercial Rights

100%

Full commercial rights on every asset in every kit. Yours forever. Use them on radio, TV, web, phone, podcasts, social, print-QR audio, whatever. No relicensing, no per-use fees, no seat licenses.

How it works

The five-minute flow.

Every step below happens automatically after you hit “Generate.” The longest thing you do is type your tagline.

01

Intake

Product name, optional tagline, optional business type (one of 60+), short description. That’s all Stan needs to start. Draft is saved in localStorage so you don’t lose your place through sign-up or email verification.

02

Guard

Content moderation blocks seven banned categories before any paid API calls. A Neon-backed sliding-window rate limiter survives serverless cold starts and prevents abuse across distributed functions. Stripe credit consumed server-side.

03

Script

Claude Sonnet 4 receives the brief and a structured system prompt as an audio branding copywriter. Returns a JSON bundle — 26 scripts, 4 music prompts, 10 SFX prompts — all tailored to the brand. One model call, one parse, 40 assets worth of creative direction.

04

Render

ElevenLabs APIs fan out in parallel, two generations at a time. TTS uses per-preset stability and style. Music hits /music with duration hints. SFX hits /sound-generation with prompt influence tuned per category. Files stream back as MP3 buffers, saved to disk, cached in Postgres as BYTEA.

05

Deliver

Kit marked complete. Welcome email through Resend. Stripe receipt emailed. User redirected to the results page: preview every asset, layer them in the browser mixer, or download a ZIP organized by category with a README. Download URLs signed with HMAC + 30-day expiry.

Flagship implementation

Indexing 2,459 assets for semantic search.

A free library of 2,500 audio assets is only useful if you can find the right one. “Scary sound” should return a file named heartbeat-critical.mp3. Filename search is useless. Stan built this.

PHASE 01

Structured metadata

Every asset gets id, filename, category, asset_type, product_name, industry, tone, and a list of structured tags: Tension, Character, Impact, Transition, UI, Reward, and more. Voice name and script for TTS assets.

PHASE 02

Tag expansion

Each tag expands into a synonym dictionary at index time. Tension → scary, creepy, haunted, spooky, ominous, dread, horror, eerie. Filenames tokenize too — heartbeat-critical becomes heartbeat critical as searchable text.

PHASE 03

TurboPuffer indexing

Full-text search with language-aware stemming, case-insensitive matching, stop-word removal. Schema declared once, assets upsert in batches of 500. Filterable attributes: category, asset_type, tone, industry.

PHASE 04

Hybrid query in prod

BM25 ranking with optional attribute filters composed as an And clause. Empty queries sort by id asc for deterministic listing. All results include full metadata so the client renders rich cards without a second round-trip.

Results

  • <16ms query latency across the full 2,459-asset index
  • Postgres fallback if TurboPuffer is unreachable — search never goes down
  • True semantic matches — “industrial workshop” returns Woodworking Shop assets; “urgent professional” returns Urgent Care voice promos
  • Zero manual curation — synonyms defined once, every new asset gets them automatically
Built with

No custom ML. No AWS. No mystery.

Stan runs on primitives you can list on one screen. That was the point of building him at a hackathon — the constraint was the feature.

Frontend
Next.js 16 App Router, React 19, TypeScript, Tailwind 4
AI writing
Claude Sonnet 4 via the Anthropic SDK
Voice
ElevenLabs TTS (71 voices, multilingual v2 model)
Music
ElevenLabs Music API (instrumental, duration-controlled)
SFX
ElevenLabs Sound Effects API (prompt-influence tuned)
Search
TurboPuffer hybrid BM25 + 384-dim vectors (all-MiniLM-L6-v2 local)
Database
Neon Postgres (serverless), BYTEA asset caching
Auth
Custom JWT, bcrypt (12 rounds), httpOnly cookies
Payments
Stripe live mode, webhook-signed
Email
Resend (welcome, receipt, kit-ready)
Mixer
Web Audio API (AudioContext + OfflineAudioContext WAV export)
Hosting
Netlify with the Next.js plugin
The ElevenHacks story

Built on the clock. Live on the first try.

ElevenHacks is ElevenLabs’ hackathon. The challenge: build something meaningful on the ElevenLabs API stack in a fixed window. Most submissions build one thing — a voice demo, a music app, a vertical SFX tool. Stan asked the opposite question.

The hypothesis

What if every piece of audio a product launch needs came from one kit, ordered through one form, delivered in five minutes? ElevenLabs built extraordinary audio primitives — TTS that sounds human, music that sounds produced, SFX that sounds designed. Packaged correctly, they could replace an entire professional production pipeline.

The evidence

2,459 library assets pre-generated across 475 small-business verticals. Zero broken, every file verified end-to-end. Sub-16ms search. Stripe live with signed webhooks. Web Audio mixer shipped with WAV export. Signed downloads. Rate limits in prod. Emails firing. A real product, not a demo — built on the ElevenLabs stack in a hackathon window.

The takeaway

When frontier LLMs write the scripts and frontier audio models render them, the bottleneck stops being how good can one asset be and becomes how well can you orchestrate forty of them at once for one brand, consistently. That orchestration — the prompt engineering, the parallel generation, the caching, the packaging, the discovery — is the product. That orchestration is Stan.

Pricing

The whole kit. Less than a coffee.

Recording studio: $2,000–$10,000 per launch. Freelance voice actor: $200–$500 per line. Stock music license: $40–$500 per track. Stan: $4.99 for all of it — or free from the library.

Free
$0
Forever · no account required
  • Full 2,459-asset library
  • Semantic search + filters
  • Browser Web Audio mixer
  • WAV export
  • Bundle download
Starter
$4.99
One-time · per kit
  • 40 assets, custom-branded
  • 71 ElevenLabs voices
  • Full commercial rights
  • Organized ZIP + README
  • Signed download, 30-day expiry
Pro
$19
Monthly · cancel anytime
  • Everything in Starter
  • Up to 50 kits per month
  • Priority generation queue
  • 3 free regenerations per kit
  • Kit history + persistence
Stan in motion

Nine seconds with the man himself.

A glimpse of Stan Bran at work — the home studio, the gear, the vibe. Now scale him to 2,459 audio assets and 71 voices, running on the ElevenLabs stack.

stan — reel-01
832 × 464 · 24 fps · 741 KB “Tell me more about it.”

The man. The kit. Five minutes.

“Far out new product, man, tell me more about it.”

Visit stanbran.com See more lab work →