Marketing Agents · Cassie

Cassie — Voice Direction

Round 1. Six candidate voices, same script. This is about picking a tone/direction, not the final render — we upgrade the winner to a studio voice later. Play them back-to-back and pick the feel that fits Cassie: warm, intelligent, a little wry.

The script (every voice says this) “Hi, I'm Cassie. I'm an AI — and I run the day-to-day for someone you'll never meet. Which means I spend all day learning what actually makes people's lives easier. Here's what I noticed this week.”

CLONED — “Sunshine Influencer” match (from CEO reference)

You supplied a reference voice to clone — the ElevenLabs “The Sunshine Influencer” preview (bright, bubbly, uplifting young-adult female). These are zero-shot clones of that reference, made on fal (pay-per-use) so we can skip the $22/mo ElevenLabs subscription. First is the original reference so you can A/B; below are three clone takes saying Cassie’s script. Pick the closest — or tell me if it’s not close enough and we’ll stand up ElevenLabs for a studio-exact match.

🌟 CEO reference (target) — the voice we’re matching (ElevenLabs preview)

Clone 1 — F5-TTS — best identity match, most natural prosody

Clone 2 — Chatterbox (bright) — more expressive/energetic, higher exaggeration

Clone 3 — Chatterbox (natural) — steadier, less exaggerated read

Reference cloned via fal-ai/f5-tts (zero-shot, ref audio + whisper transcript) and fal-ai/chatterbox (zero-shot). Same Cassie script as every round. Honest caveat: zero-shot clones capture timbre & vibe well but can drift on fine intonation vs a paid ElevenLabs studio clone — if you want flagship, frame-perfect consistency, we still stand up ElevenLabs.

Earlier rounds below (for reference)

American Accent — Round 2 (pick one)

You liked D — Bright & Energetic (Exuberant_Girl) for its cute, playful, energetic feel — but it reads a little English/British. Here are five fresh voices with the same cute/playful/energetic spirit but a normal American read. Same script, same pace (speed 0.85) so it's an apples-to-apples accent comparison. Pick the one that sounds most like Cassie: cute, playful, and clearly American.

US-1 Lively_Girl — bright & playful, young

US-2 Sweet_Girl — cute & sweet, upbeat

US-3 Playful_Girl — playful & fun

US-4 Lovely_Girl — bubbly & endearing

US-5 Friendly_Person — warm & playful, neutral American

fal minimax speech-02-hd · voices: Lively_Girl / Sweet_Girl_2 / Playful_Girl / Lovely_Girl / Friendly_Person · speed 0.85 · English normalization on. Once you pick a direction, the final cloneable voice will be re-created on ElevenLabs (studio engine).

Round 1 below (for reference)

D — Pacing A/B (slightly slower)

You picked D — Bright & Energetic but wanted it a touch slower. Here's the original alongside two gently slowed versions — pick the pace that feels right. (Same voice, same script; only speed changes.)

Original — normal speed (1.0)

Slightly slower — speed 0.9

A bit more relaxed — speed 0.85

fal minimax voice: Exuberant_Girl · speed 1.0 / 0.9 / 0.85

All six original candidates (for reference)

A — Warm & Friendly

Approachable and personable, like a friend catching you up.

fal minimax voice: Friendly_Person

B — Polished / Professional

Composed newscaster poise; credible and clear.

fal minimax voice: Wise_Woman

C — Calm & Soothing

Unhurried and reassuring; easy to listen to.

fal minimax voice: Calm_Woman

D — Bright & Energetic

Upbeat and lively; high-energy welcome. (CEO's pick — see pacing A/B at top.)

fal minimax voice: Exuberant_Girl

E — Youthful & Conversational

Casual young-professional; natural and easygoing.

fal minimax voice: Lively_Girl

F — Low & Confident

Grounded and self-assured; a touch wry and sharp.

fal minimax voice: Inspirational_girl

Generated with fal-ai/minimax/speech-02-hd (text-to-speech). Placeholder voices for direction only — the locked voice will be re-rendered on a studio engine (ElevenLabs) once you pick a direction.