What this is — and what it isn’t
This is not a séance, a chatbot, or a way to talk to someone who is gone. There is no consciousness on the other side — no memory of your life together, no answers you did not write. What vocalove does is narrower and, for many families, enough: clone a voice from a real recording you have permission to use, then speak one short line you choose — their cadence on your words.
That honesty matters. People search for ways to talk to a loved one again when grief needs a voice, not a product demo. We say the boundary up front because misleading someone in that moment breaks trust faster than any feature list builds it. You type a question. The tool renders a brief answer in their cloned voice — one take, a few seconds — with the question shown as a subtitle. It feels like a reply; it is not a conversation.
If you need a longer tribute or a talking portrait for a service, use our memorial video. If you only need the voice as audio, you can skip the photo and download the clip. This page is for the single sentence you still need to hear — once, not on loop.
How it works
Three steps. No chat thread. No typing on their behalf beyond the one answer line the system speaks.
Step 1
Record or upload a voice sample
Three to thirty seconds of clear speech you have permission to use — a voicemail, a toast, a family clip. Quiet room, one speaker. This sample teaches the clone how they sounded; it is not the answer itself.
Step 2
Type one question
One text box, one question — about a hundred characters. “Are you okay?” “Did I do right by you?” “What would you tell the kids?” The question appears as a subtitle on the finished clip; it is not spoken aloud.
Step 3
Generate a four-to-six-second answer
The tool writes a short compassionate answer, clones it in their voice, and renders one video (portrait optional). One question, one answer, one charge. Ask again anytime — each new question is a new generation.
One question, one answer
We deliberately did not build a chat thread. Emotionally, one question and one pause feels like speaking to someone; a scrolling back-and-forth feels like customer support. Grief does not need a typing indicator.
Technically, every extra line is a full clone-and-render pass — talking photo, timing, export. A multi-turn “conversation” would multiply cost and failure points without making the moment more true. One take keeps the answer short enough that mouth movement stays believable — four to six seconds, roughly one or two sentences.
You can still ask again tomorrow with a different question. Each ask creates a new video and deducts credits again. The UI says so plainly: this is not a session, it is a keepsake per line. Many families generate two or three over a week, privately, and stop when the sentence they needed has been heard.
Questions people ask first
Tap one to fill the box, or write your own. These are starting points — not scripts the tool repeats blindly. The spoken answer is always a fresh short line matched to your question.
If you don’t have a recording
Without a sample, cloning cannot run — we will not substitute a stranger’s voice. Four places families often find one:
Voice messages on a family phone
Search the camera roll and message apps for audio they sent — birthday wishes, travel check-ins, a quick “love you” before bed. Export the file; three seconds of clear speech can be enough.
Voicemail
Carrier voicemail can often be saved or forwarded before it expires. Ask the phone account holder for permission; a short greeting is a common last sample.
Audio from old videos
Wedding speeches, holiday clips, a child’s recital in the background — extract audio in any editor, or play the clip in a quiet room and record a clean capture if rights allow.
Home movies and digitized tapes
Transfer services return files you can scrub for dialogue. Pick a moment with one face and one voice, not crowd noise.
If nothing turns up yet, you can try a built-in voice to test the flow — but the moment that lands is still their cloned voice, not a generic narrator.
Their voice, not a stock narrator
Built-in voices are fine for a quick test or when you have no sample yet. For “I love you” at midnight, clone mode is the point — upload a sample so guests hear a timbre they recognize, not a stock narrator.
The voice cloning tool on the homepage also runs in your browser: sample, script, generate. Use it when you want to preview a line before you pay for a full export, or when you need audio without a portrait. For a longer comparison of browser vs desktop clone software, read our voice cloning software guide.
Permission & privacy
Only clone voices you own or have explicit permission to use — yourself, a partner, a parent, a relative the family agrees on for a memorial. Do not upload celebrities, coworkers, or strangers. Public figures and non-consensual clones violate our Acceptable Use Policy and may be removed.
Uploads are used to generate your output — not sold for ads or unrelated model training. Delete creations from your account when you want; contact us to remove your account and associated files. Preview privately with relatives before you share outside the family.
What it costs
100 credits per answer — roughly a dollar at pack rates — for one four-to-six-second video (or audio if you skip the portrait). One-time credit packs, no subscription. This page does not include a free trial; preview clone quality on the homepage if you need to hear a match before you spend.
Each new question is a new generation and a new charge. Credits never expire after purchase. See credits & pricing for Starter, Creator, and Pro packs.