AI VOICE DETECTOR · FREE PLAN
Is This Voice Real or AI-Cloned?
Voice cloning tools can imitate a person from short recordings. The AnyFact voice detector scores each segment and checks the message for scam patterns.
Open in beta · 10 free credits a month · Files deleted after 30 days
AI Voice Detector
Open in betaWhat you need
Upload a voice note or recording
MP3, WAV, M4A or OGG (WhatsApp voice notes), up to 5 minutes.
At a glance
Per segment
Each segment of the recording is scored, and the transcript is checked for scam patterns.
AnyFact method
3 levels
Probably human, uncertain or probably AI — never a bare percentage — with the reasons and the limits of the method.
AnyFact method
10 credits
Free every month with an account, no credit card. Uploaded files are deleted after 30 days.
AnyFact free plan
Why a familiar voice is no longer proof
Voice cloning has become cheap and accessible. A voice note that sounds like a relative, a manager or a bank adviser is not, on its own, proof of who sent it.
Cloned-voice scams usually follow a script: urgency, secrecy, a request for money or for changed bank details. The words matter as much as the sound.
The AnyFact voice detector scores each segment with an open-source anti-spoofing model, transcribes the message and flags those scam patterns. Whatever the result, the safest check remains to call the person back on a number you already know.
Does this sound familiar?
- A relative sends an urgent voice note asking for money.
- A “manager” asks for a transfer by voice message.
- A voice message asks you to keep a request secret.
Here's what we check
Upload the audio
MP3, WAV, M4A or OGG (WhatsApp voice notes), up to 5 minutes.
Segment scoring and transcript
Suspicious segments are marked on the waveform; the transcript is checked for scam patterns.
A verdict in three levels
Probably human, uncertain or probably AI — never a bare percentage — with a short explanation and the limits of the method.
How to read the result
The signals point to a person. It is not a guarantee: a careful generator or heavy editing can leave few traces.
The signals disagree or are too weak. We say so rather than guess, and list what was found.
The signals point to generation. It is an indication, not proof — check the context before acting.
Sample layout. Real reports name the signals that were checked and the limits of the method.
How to check a voice message with the AI voice detector
Upload the original recording and read both parts of the result: the sound and the words. While it runs, verify the request through another channel.
- 01Save the original file from the app that received it. Re-recording a message played through a speaker degrades it.
- 02Upload an MP3, WAV, M4A, OGG, WebM or OPUS file, up to 25 MB and 5 minutes.
- 03Read the segment scores: each part of the recording is scored for synthetic speech.
- 04Read the transcript and the scam patterns flagged in it.
- 05Whatever the result, call the person back on a number you already know before you act.
A voice analysis uses 2 credits.
What does an AI voice checker listen for?
It checks how the voice sounds and what the message says, because cloned-voice scams depend on both.
| Check | How it works | What it shows |
|---|---|---|
| Provenance | C2PA content credentials and metadata are read first | A known AI generator, if the file names one |
| Synthetic speech | An open-source anti-spoofing model (wav2vec2-based) scores each segment | Which parts of the recording sound generated |
| Transcript | Whisper transcribes the speech | The exact words, so you can read them calmly |
| Scam patterns | The transcript is checked for urgency, secrecy, money requests and changed bank details | Pressure tactics, whatever the voice |
The transcript is useful on its own. Reading the words on a screen, away from the emotion of a familiar voice, makes pressure tactics much easier to notice.
How to read a voice clone detector result
Read the voice score and the scam patterns together. Their combination tells you more than either one alone.
| Voice | Scam patterns | What it suggests |
|---|---|---|
| Probably AI | Flagged | Treat the message as high risk. Do not pay or share anything; call back on a known number. |
| Probably human | Flagged | A real person can still be scamming you, or a clone went undetected. Verify before acting. |
| Probably AI | None | Possibly a legitimate synthetic voice, such as a voiceover or an accessibility tool. Check who sent it. |
| Uncertain | Any | Often a noisy or heavily compressed recording. Rely on a callback, not on the score. |
When is deepfake audio detection harder?
Detection is harder whenever the recording has been degraded or mixed with other sounds. Expect more uncertain results with:
- Background noise, music or several people talking at once.
- Phone-line quality or heavy compression from messaging apps.
- Audio re-recorded from a speaker instead of exported from the app.
- Very short clips with only a few words.
- Recordings over 5 minutes: trim them to the part that contains the request.
Our guide on how to tell if a voice is AI describes what to listen for yourself. If a voice note comes with video, the deepfake video detection page checks both the image and the soundtrack.
How to protect yourself from voice clone scams beyond detection
Simple habits stop most voice-based pressure, whether the voice is cloned or not.
- 01Agree on a code word with close family, to ask for when a call feels wrong.
- 02Hang up and call back on a number you already have, never the one in the message.
- 03Never change bank details or send money on the strength of a voice message alone.
- 04Slow down when you are told to hurry or to keep a request secret.
- 05Tell your bank or your company’s finance team if you think you were targeted.
Our page on voice clone scams covers the scenarios families and finance teams face most often.
Before you check
Is the result proof that a content was made by AI?▼
No. It is an indication, not proof. Detectors can be wrong in both directions, especially on compressed, cropped, edited or very short content. Use the result to decide what to check next, not as a final judgement on anyone.
Why no percentage?▼
A single percentage suggests a precision no detector has. You get one of three levels — probably human, uncertain or probably AI — with the signals behind it and the limits of the method.
How does the analysis work?▼
When the content carries provenance data, such as C2PA content credentials or metadata, we read it first. Then an open-source detector scores the content. When the score is borderline, Claude reviews it as a second opinion, and the result is explained in plain words.
What happens to my files?▼
They are processed to run your analysis — including, for borderline cases, a review by Claude, Anthropic’s model — and deleted after 30 days.
What does it cost?▼
There is a free plan: a free account gets 10 credits every month, with no credit card. Pro ($15 a month) gives 1,000 credits a month; Business ($49 a month) gives unlimited Business Checks and AI checks under a fair-use limit.
Can the AI voice detector check a phone call while it is happening?▼
No. It analyses recordings you upload, not live calls. If you can record or save the message, you can check it afterwards.
Does it work on voice notes from messaging apps?▼
Yes, as long as you export the file. Voice notes are usually saved in formats such as OPUS, OGG or M4A, which are accepted, up to 25 MB and 5 minutes.
Can it tell whose voice was cloned?▼
No. It estimates whether the speech sounds synthetic, not whose voice it imitates. Only the real person, reached on a number you trust, can confirm whether they sent the message.
What if my recording is longer than 5 minutes?▼
Trim it to the part that matters, usually the section where the request is made, and upload that part.
Test a voice message free — no account needed for your first analysis
What you need
Upload a voice note or recording
Free plan — 10 credits every month, no credit card.
Each result is an indication, not proof.
Files deleted after 30 days.
You might also want to check
Deepfake Detector: Is This Video, Photo or Voice Real?
Deepfake detection for videos, photos and voice messages: face swaps, generated clips and cloned voices, checked with open-source detectors and explained. Free plan.
Is This Video Real or AI-Generated?
Deepfake video detection and AI video detector: frames, faces and soundtrack are checked separately and suspicious moments are marked on a timeline. First test free.
Was This Text Written by AI?
Free AI text detector and AI writing checker: paste a text to see which passages were probably written by AI, with an explained verdict. English, French, Arabic.