FakeYou
FakeYou is a US platform for AI audio and video generation, built around a community library of over 3,500 character voices. It bundles text to speech, voice conversion, lip sync and video tools, with a permanent free tier.
What is FakeYou?
FakeYou is a web platform published by Learning Machines, Inc, trading as Storyteller, that turns text, audio and images into AI-generated speech and video. Its home page headline sums the offer up as AI Music, Text to Speech and Voice to Voice, and the product grew out of a voice deepfake engine for famous and fictional characters.
The centerpiece is the text-to-speech library: more than 3,500 character voices, most of them trained and contributed by members rather than by the company. Voices can be filtered by language, community rating or featured status, and each one carries ratings and comments; a leaderboard and public profiles complete the community layer. Anyone can train, upload and share their own voice model, which explains both the size of the catalog and its uneven quality.
Around that library sits a broader tool set. The AI Tools block lists Text to Speech, Voice to Voice (speak in another voice), Voice Designer (create your own AI voice), Text to Image, Lipsync, Live Portrait (transfer motion and audio between portraits) and Video Style Transfer. The side navigation splits the same catalog into Speech Generation and Video Generation, the latter adding Face Animator, Video Motion Capture and a Video Workflow. Several open-source engines are exposed by name inside the app: F5-TTS for zero-shot text to speech, Seed-VC for zero-shot voice conversion from a one-to-thirty-second reference, CogVideo for text and image to video, and Hallo 2 for audio-driven portrait animation.
Basic generation works without an account, while a free account keeps your history. Jobs run through a visible queue whose priority depends on your plan, voices cover 17 languages and the interface is translated into 14. A full API is reserved for the Pro and Elite tiers. The same team also builds Storyteller Studio, a 3D scene tool in beta, and ArtCraft, an open-source desktop and web generator promoted on the FakeYou site.
The publisher frames the service carefully: the terms call it a research technology for fun, watermark generated audio and forbid commercial use except with specially denoted commercial voices.
What it does
- Generate speech from text with one of 3,500+ character voices
- Turn your own recorded voice into another voice, spoken or sung
- Clone a voice from your own audio files with Voice Designer
- Animate a still portrait or lip-sync a face to an audio track
- Apply an AI style to a short video clip
- Generate images from a text prompt
- Upload and share your own voice models with the community
When to use FakeYou / When not to
A quick filter to help you decide if FakeYou is the right fit.
When to use FakeYou
- Social media creators producing YouTube, TikTok or Shorts clips with recognizable character voices
- Meme, parody and fan-content makers looking for a large ready-made voice catalog
- Musicians and cover artists converting sung vocals through Voice to Voice
- Game, animation and comic teams prototyping character voices before casting
- Voice-AI hobbyists who want to train, upload and share their own models
When not to use FakeYou
- Businesses needing commercially usable output: the terms ban commercial use outside specially denoted voices
- Regulated industries, since HIPAA, FISMA and GLBA workloads are explicitly excluded by the terms
- Enterprise buyers requiring a DPA, a named subprocessor list or SOC 2 / ISO 27001 evidence
- European organizations needing EU data residency: hosting is US-only and no Article 27 representative is designated
- Anyone producing long-form audio or video: free audio stops at 12 seconds and style transfer at 3 seconds
How to use FakeYou
A typical end-to-end flow, from setup to results.
- Open the site and pick a tool: Text to Speech, Voice to Voice, Voice Designer, Text to Image, Lipsync, Live Portrait or Video Style Transfer
- Browse the catalog of 3,500+ voices, using search, language filters, community ratings and featured voices to find a model
- Type or paste your text, keeping punctuation clean and taking the character's speaking style into account
- Click Speak, then listen, download or share the result; create a free account if you want to keep your generation history
- For Voice to Voice, choose a target voice model, record or upload your audio, set the pitch shift (up to 36 semitones) and the pitch estimation method, then click Convert
- For Seed-VC, upload a 1 to 30 second reference of the target voice, add your source audio and click Convert Speech; for F5-TTS, add a reference clip, type your text and click Generate Speech
- For Live Portrait, pick a sharp portrait photo and a driving video (preset or uploaded), click Animate and download the clip
- For Video Style Transfer, upload a video (up to 7 seconds on Pro and Elite), choose a style, set intensity from 0 to 100%, optionally enable Face Detailer, Cinematic or Upscaler and request up to three variants
- Read the dedicated TTS guide to fix common issues: closing punctuation against babbling, deliberate misspellings for pronunciation, emotional conditioning sentences to delete afterwards and Arpabet annotation on supported models
- Contribute or upgrade: upload your own TTS models, voice models, LoRA or Stable Diffusion weights and lipsync templates, request a custom clone through the dedicated page, or subscribe through Stripe Checkout and manage the plan from the Stripe portal
Pros & Cons
Pros
- Unusually large voice catalog, with more than 3,500 character models
- Permanent free tier offering unlimited generation, within short duration limits
- Wide range of tools gathered in one place: voice, video, image and motion capture
- Recent open-source engines exposed under their own names (F5-TTS, Seed-VC, CogVideo, Hallo 2)
- Active community with Discord, leaderboard, ratings, comments and Patreon, and the ability for anyone to train and share models
- Detailed TTS prompting guide that is candid about the weaknesses of the models
- Public pricing with no quote required, full API from the Pro tier, self-service cancellation through the Stripe portal, 17 voice languages and an interface in 14 languages
Cons
- Commercial use is forbidden by the terms outside specially denoted voices, which rules out most professional work
- Very short generation limits: 12 seconds of TTS audio on the free tier, 30 seconds on Plus, 1 minute on Pro, 2 minutes on Elite, and video style transfer capped at 3 seconds free and 7 seconds on Pro and Elite
- Uneven voice quality, acknowledged by the site itself, with older models producing incoherent babbling
- Light compliance posture: no postal address published anywhere, no DPA, no named subprocessors, no certification, and a privacy policy unchanged since 18 November 2022 while the product moved on
- The site is entirely JavaScript-rendered, so nothing is readable without a modern browser, and the API documentation URL published in the footer (docs.fakeyou.com) returns a 'Page not found' host page
- Audio is watermarked, video watermark removal is Pro and Elite only, ads are shown on the free tier and queue priority there is the slowest
- Very broad content license granted to the publisher, no documented opt-out from model training, an age contradiction between the terms (13) and the privacy policy (18), and a cancellation path described in the terms as going through an email contact
Pricing & Plans
FakeYou offers a permanent free plan, called Starter, at USD 0. The lowest paid entry point is the Plus plan at USD 12 per month, followed by Pro at USD 25 per month and Elite at USD 40 per month. Payment and billing are handled by Stripe, through Checkout and the customer portal. No annual pricing, no currency other than the US dollar and no time-limited free trial are advertised: the free tier plays that role. Sales taxes may be added where applicable, and cancellation takes effect at the end of the paid period in progress.
- 'Slowest' processing priority
- unlimited TTS generation up to 12 seconds of audio
- Voice to Voice up to 3 minutes
- access to Lipsync and Live Portrait
- Video Style Transfer with 3-second generations
- TTS text length capped at 1
- 024 characters
- 'Normal' priority
- unlimited TTS up to 30 seconds
- Voice to Voice up to 4 minutes
- all ads removed
- access to Lipsync
- Live Portrait and 3-second Style Transfer
- TTS text length 2
- 048 characters
- 'Faster' priority
- unlimited TTS up to 1 minute with private model uploads
- Voice to Voice up to 5 minutes with private models
- no ads
- Live Portrait with private videos and watermark removal
- 7-second Style Transfer with private videos
- watermark removal and faster
- higher-quality renders
- 'Fastest' priority
- unlimited TTS up to 2 minutes with private model uploading and sharing
- unlimited Voice to Voice audio with shareable private models
- no ads
- Lipsync
- Live Portrait and Style Transfer at Pro level
- full API access
- FakeYou commercial voices
- All three paid tiers also advertise '+ Future feature updates'
Data, GDPR & hosting
A consolidated view of how FakeYou handles your data.
GDPR overview
FakeYou's privacy policy, last updated on 18 November 2022, names Learning Machines, Inc (dba Storyteller) as controller and includes a section for people located in the EU and the UK that cites the GDPR and the UK GDPR. It lists legal bases (consent, contract performance, legitimate interests, legal obligations and vital interests) and announces access, rectification, erasure, restriction, portability and objection rights for the EEA, the UK and Canada, plus withdrawal of consent at any time. Requests go through a form or support@storyteller.io, with links to the EU supervisory authorities and the Swiss federal commissioner. A separate CCPA section covers California residents. Missing, however: no Article 27 EU representative, no named data protection officer, no DPA offered, no named subprocessor list and no certification such as SOC 2 or ISO 27001. Hosting and processing are contractually located in the United States, and compliance is never claimed in so many words.
Who owns the data?
FakeYou's terms state that Learning Machines, Inc claims no ownership over user Contributions: you keep full ownership and the associated intellectual property rights. In exchange you grant the publisher a non-exclusive, unlimited, irrevocable, perpetual, worldwide, transferable, royalty-free and sub-licensable license to host, copy, sell, resell, publish, broadcast, translate, extract from and adapt your content, for commercial or advertising purposes, explicitly covering your image and voice, with a waiver of moral rights. Submissions such as questions, comments, suggestions and feedback become the publisher's exclusive property without compensation. Contributions can be edited, recategorized or removed at any time, and rights holders, original speakers or voice actors may request takedowns at copyright@storyteller.io.
Reuse rights
FakeYou collects account data (email addresses, usernames, passwords, contact preferences, names, card numbers) and automatic data (IP address, browser and device characteristics, operating system, language preferences, referring URLs, country, location and usage logs). Stated purposes are account creation and management, order fulfilment, feedback requests, marketing communications, security and fraud prevention, usage-trend analysis and protection of vital interests. The publisher says it processes no sensitive data, receives none from third parties and has sold no personal information under the CCPA in the past twelve months. Payments are handled and stored entirely by Stripe, while Google Analytics and a product-analytics tool of the PostHog type are present in the code. Recipients are described only by category: cloud computing, data analytics, payment processors, performance monitoring, and registration and authentication services. Social sign-in shares profile data from the provider (name, email, friend list, picture). The privacy policy states that uploaded audio is not analyzed to identify it, sell it or use it in any way other than to provide and improve the service, a wording that never settles whether customer data feeds model training. Do-Not-Track signals are not honoured. On the output side, users may reuse their own generations without asking permission, but the terms restrict that reuse to non-commercial purposes unless the voice is specially denoted as commercial.
Data retention & training
Hosting summary
The terms state explicitly that the site is hosted in the United States, and no alternative hosting country or region is offered. Any user located elsewhere accepts, by continuing to use the site, that their data is transferred to and processed in the United States; there is no European data-residency option and no standard contractual clauses are mentioned. No infrastructure provider is named in the public documents: the privacy policy only refers to the category 'Cloud Computing Services'. Payment data is handled and stored by Stripe. The domain's public IP address (172.66.146.28) belongs to Cloudflare, Inc. (AS13335) and is anycast, so it identifies a CDN point of presence rather than a storage location and should not be read as evidence of where data physically sits. Nothing in the public documents describes backup locations, replication or any jurisdiction other than the United States.
Where FakeYou works
Country-level availability.
Not available in
Things to keep in mind
Risks and trade-offs to weigh before adopting FakeYou.
- Commercial use is forbidden by default: publishing FakeYou output in monetized or client work breaches the terms unless the voice is specially denoted as commercial
- The terms require every published usage to be labeled as "deep fake", and state that the publisher does not condone impersonation, deception, slurs, abuse or mistreatment of any individual or group; realistic voice cloning makes those harms easy, and the consequences fall on you
- Character and celebrity voices raise publicity, personality and neighbouring rights issues that the platform does not cover: it places responsibility on the user and requires written consent from every identifiable person appearing in uploaded content
- The license granted on contributions is unlimited, irrevocable, perpetual, worldwide and transferable, explicitly covers your image and voice, includes a waiver of moral rights, and any suggestion or feedback you send becomes the publisher's exclusive property
- Legal recourse is narrow: mandatory arbitration, class-action waiver, Cobb County jurisdiction in Georgia, a one-year contractual limitation period and liability capped at what you paid over the previous six months
- Everything is processed in the United States, a transfer that any European user accepts simply by continuing to use the site, and there is no commitment against using your data for model training
- The legal documents are dated 18 November 2022 while the toolset has changed a lot, no postal address is published anywhere, the only contact is an email on a neighbouring domain (storyteller.io), and the terms and privacy policy disagree on the minimum age, 13 against 18
Setup & Integrations
Technical difficulty
Low for everyday use. FakeYou is entirely web-based, with nothing to install. Basic generation works without an account, and a free account (email and password, or social sign-in) simply stores your history. Text to speech takes three steps: pick a voice, type your text, click Speak. The real learning curve is quality: the official guide documents punctuation tricks, deliberate misspellings and Arpabet annotation, and the video tools require well-framed source media. Waiting times depend on the queue and your plan. API access is Pro and Elite only, and its documentation is currently unavailable.
Deployment
Supported languages
Behind FakeYou
Fundraising
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
How many voices does FakeYou offer?
Can I use FakeYou output commercially?
Is there a free plan?
How much do the paid plans cost?
Does FakeYou have an API?
Can I clone my own voice?
Is the generated audio watermarked?
What is the minimum age to use FakeYou?
Where is my data hosted?
Which languages are supported?
Should you pick FakeYou?
FakeYou started as a voice deepfake engine for famous and fictional characters and has grown into a broader creative suite covering speech, video and images. Its defining strength is the community catalog of more than 3,500 character voices, a scale that is hard to find elsewhere, backed by ratings, comments, a leaderboard and the ability for any member to train and publish a model.
Access is genuinely free at the Starter tier, and the paid plans are modest and publicly listed: USD 12, 25 and 40 per month, with no quote and no sales call. Recent open-source engines (F5-TTS, Seed-VC, CogVideo, Hallo 2) are exposed under their own names.
The limits are just as clear. The terms forbid commercial use outside voices specially denoted as commercial, which disqualifies most professional projects before anything else is considered. Generation durations are short at every tier: 12 seconds of free audio, two minutes at the top plan, three to seven seconds of video style transfer. Compliance is light: no DPA, no named subprocessors, no certification, US-only hosting, no Article 27 EU representative, and a privacy policy untouched since November 2022 even though the tool catalog has changed considerably. The site also contradicts itself on the minimum age (13 in the terms, 18 in the privacy policy), and the API documentation URL it publishes is offline.
The publisher, Learning Machines, Inc trading as Storyteller, is a US company whose Atlanta team is now also promoting other products, Storyteller Studio and ArtCraft, which is worth knowing if you plan to build a workflow on FakeYou.
In short, FakeYou is an excellent creative and community playground, best suited to personal, experimental and fan projects, and to be recommended with real caution for professional production.
- Choosing a selection results in a full page refresh.
- Opens in a new window.