FAQ
Questions we hear in the booth
Straight answers about how VoiceSynth works as a generative-AI voice studio — engagement models, budgets, tools, consent, and the lines we will not cross.
Is VoiceSynthAI a free text-to-speech app, or can you clone anyone's voice?
No. We are a generative-AI voice studio — a creative services firm that produces voiceover, narration and localization from consented, licensed voices with human direction. We are not a self-serve TTS app or SaaS, and not a course. We do NOT clone real people's voices without explicit written consent and will not create impersonation, scam or deceptive audio. We clarify usage rights on deliverables; "Synth" means voice synthesis; the .pro TLD is branding only.
How do engagements work — project, retainer, or per deliverable?
Most clients start with a scoped CAD project: script in, cast and directed voiceover out, with defined revision rounds. Retainers suit teams with ongoing narration, localization, or in-product prompt updates — you reserve studio capacity monthly and we queue deliverables against an agreed scope. Per-deliverable pricing works for one-off spots or single-language pickups. We will recommend the model that matches your turnaround and budget honestly.
What are typical CAD budgets?
Short English narration (under five minutes) commonly falls between C$800 and C$2,500 including two revision rounds. Multi-language dubbing is priced per language, often C$1,200–C$4,500 depending on script complexity and lip-sync requirements. Custom consented voice design starts around C$3,500 and scales with language coverage and review cycles. Retainers begin near C$2,800 per month for modest ongoing deliverables. We quote before work begins — no credit-metered SaaS surprises.
What turnaround should we plan for?
Standard narration projects often ship in three to five business days after casting approval. Multi-language dubbing and custom voice design need longer timelines — sometimes four to eight weeks. Rush work is possible when scope is tight and studio capacity allows; we will be direct when a deadline is not realistic without compromising human review.
Which speech synthesis tools and models do you use?
We select text-to-speech, speech-to-speech, and voice design tools based on language, voice character, and licensing fit — not a single vendor religion. Model and voice selection happens in casting, with neural voices from licensed libraries or consented custom builds. Tools change; our standard does not: every generative-AI output is reviewed by a human director before delivery.
Who owns the audio and what usage rights apply?
Ownership and licensing depend on the statement of work. Typically clients receive broad usage rights for agreed channels — digital, broadcast, in-product, e-learning — within specified territories and durations. We document restrictions clearly. Custom voices may carry additional publicity or talent obligations. We do not provide legal advice; clients should have counsel review high-stakes licensing questions.
How do you handle voice consent and likeness rights?
Licensed library voices come with vendor terms we pass through in plain language. Custom or consented voice cloning requires written consent from the voice talent and clarity on publicity rights. Clients supplying reference audio must confirm they have permission for studio use. We maintain consent records aligned with PIPEDA. We refuse projects that ask us to imitate real people without documented permission.
How is personal information handled under PIPEDA?
Contact details, brief materials, and voice samples are collected only for stated purposes — quoting, producing, and delivering your project. See our Privacy Policy for retention, access rights, and how to reach the Office of the Privacy Commissioner of Canada. Form consent is required and never pre-checked.
What do you not do?
We do not operate a self-serve TTS subscription, stock-audio marketplace, clone-anyone tool, deepfake or scam audio service, AI-audio course, NFT mint, get-rich-quick scheme, or wellness meditation product. We do not produce explicit or unlawful content, impersonation audio, or deceptive speech presented as a real person's genuine statement. We are a Toronto voice studio with human direction and responsible AI practice.
What file formats do you deliver?
Master WAV is standard, with MP3 or platform-specific formats on request. Loudness targets follow your channel — broadcast, streaming, or in-product. We can deliver split stems, pronunciation guides, and style notes for future revision rounds when scoped.
Still have a question?
Send a brief outline — we will reply with scope and next steps.
Start a voice brief