TopMediai is a browser-based AI studio that generates video, music and voiceovers from one account and one shared credit pool, aggregating third-party video models such as Seedance, Veo and Kling alongside its own music and 3,200+ voice library.
Used this tool? Rate it
Used this tool? Rate it
TopMediai is a browser-based AI content studio that generates video, music and voiceovers from a single account. Its own description of itself is direct: an all-in-one AI content creation studio for video, music and voice production, built so that you do not need separate tools for visuals, voiceovers, background music, dubbing and audio creation.
The structural fact that matters most is how it is organised. Unlike vendors that scatter capabilities across a family of separately branded sites, everything here lives under one domain, behind one login, and — critically — draws on one shared credit pool. The pricing page states this as its headline: one credit system, all creative tools, with credits usable across the entire platform. Whether you spend a credit on a song, a video clip, a voice clone or a minute of dubbing is your decision rather than a licensing boundary.
The second structural fact is that TopMediai is, on the video side, substantially an aggregator. Its published list of supported models includes Seedance 2.5, GPT-Image-2, Veo 3, Sora 2, Vidu 2.0, PixVerse, Wan 2.5, KlingAI, Nano Banana, Midjourney, Nano Banana Pro, Seedream, Grok and Hailuo. That is a meaningful thing to understand before you buy: for text-to-video and image-to-video work you are largely paying for unified access, credits and a workflow wrapper around models built elsewhere, not for a proprietary video model. On the audio side the picture is different — the music generator and the voice library appear to be the platform's own, and that is where its distinctive value sits.
Who operates it. The terms page states that topmediai.com is operated by Cleverguard Technology Co., Limited, and lists three offices with registration numbers: Hong Kong (Star House, Tsim Sha Tsui, registration 72039776), London (36 Gerrard Street, registration 13291377) and Denver, Colorado (1500 N Grant St, registration 20241438604). The privacy policy independently identifies the same entity as operating under the TopMediai name. There is a discrepancy worth flagging: the third-party directory GoodFirms lists the developer as Shenzhen iMyfone Technology Co. Ltd, founded 2023, China. We report both rather than reconciling them; the terms page is first-party and therefore the stronger source, but the divergence is real.
Self-reported scale figures on the homepage are 3.5M+ creators, 10M+ AI videos, 3M+ songs and 8M+ voice creations. These are vendor claims with no stated independent audit.
The music side is the platform's most developed original capability. It includes an AI music generator, AI Cover for reworking songs with different vocal styles, a Singing Photo Maker that animates portraits into singing clips, Audio to MIDI conversion, a Stem Splitter, an Instrument Separator and a Music Hook tool for short catchy segments.
The Audio to MIDI and stem separation tools deserve a note because they are the ones that serve people who already make music, rather than people who want music made for them. Splitting a track into stems and extracting MIDI are conventional production tasks, and their presence suggests the platform is trying to sit alongside a DAW rather than only replace one.
The voice stack covers text to speech, speech to speech, AI voice cloning, a voice library, a voiceover studio, an AI voice changer and a voice enhancer. The published library figure is 3,200+ voices across 190+ languages and accents, which the company positions for narration, character voices and multilingual dubbing.
Speech to speech is worth distinguishing from text to speech: rather than reading a script, it transfers your delivery — timing, emphasis, emotion — onto a different voice. For anyone who has fought with TTS prosody, that is often the more useful of the two.
Video entry points include image to video, text to video, video to video, script to video, music to video, AI lip sync and AI video effects. There is also a long tail of format-specific generators the site promotes heavily — UGC video, AI influencer, Minecraft parkour, cat videos, ASMR, Subway Surfers style clips, news video — which tells you plainly that a large share of the intended audience is short-form social creators working in recognisable viral formats.
Five documented APIs are offered: AI Music Generator, AI Song Cover, Text to Speech, Voice Cloning and Voice Changer. Note what is on that list and what is not — the APIs cover the audio stack, which is consistent with audio being the platform's own technology and video being aggregated.
TopMediai describes a four-step pipeline: create AI music and soundtracks, generate AI videos, add AI voiceovers and dubbing, then polish and share. The ordering is deliberate and slightly unusual — music first, video second — which reflects a genuine use case where the track drives the visuals rather than the reverse.
Short-form social video with original music. This is the core case. Producing Shorts, Reels and TikToks normally means sourcing a track, licensing it, generating a voiceover somewhere else and assembling the result. Doing all three in one place with one credit balance is the actual product proposition.
Faceless content channels. The platform explicitly targets creators who want to produce content without filming themselves or recording their own voice. The dedicated generators for Minecraft parkour, ASMR, cat videos and similar formats are built for exactly this segment.
Multilingual localisation. With a library spanning 190+ languages and accents plus a video translator, dubbing an existing video into additional languages is a first-class workflow rather than an afterthought.
Product promos and e-commerce creative. Turning product images and selling points into promotional video with a soundtrack and a voiceover is a repeatable, high-volume task that suits credit-based pricing well.
Original songs and music videos. For independent artists the combination of song generation and the Singing Photo Maker or music-to-video path produces a releasable audio-visual package from a single idea.
Story and narrative video. Multi-character AI voices plus scene-based visuals and a custom score is a genuinely awkward thing to assemble across separate tools, and it is one of the cases where the unified-credit design earns its keep.
Treat the credit table as an upper bound, not a promise. Official conversions are stated as "up to" figures — for 1,500 credits, up to 150 songs, up to 83 videos, up to 300,000 characters of text to speech, up to 15 voice clones, up to 50 song covers. Retries, failed generations and longer outputs all consume from the same pool.
Batch your voice cloning. Clone counts are among the more constrained allowances (15 per 1,500 credits), so establish the voices you need rather than experimenting one at a time.
Cancel deliberately and early if you are trialling. Given weekly billing on the entry tier and a refund policy that excludes promotional purchases, an unwanted renewal is easier to trigger and harder to reverse here than on a typical monthly SaaS product.
Do not rely on the aggregated model roster staying fixed. Which third-party video models are available is a commercial arrangement, not a product guarantee. If a specific model is essential to your work, that dependency is worth acknowledging.
Prefer generic voices over character or celebrity ones for anything you publish. The reasoning is in the next section, and it is a legal reason rather than an aesthetic one.
Keep prompts and settings alongside outputs. Credit-metered generation makes reproducing a good result expensive if you have to rediscover how you got it.
High-volume short-form creators get the clearest benefit. If you publish daily and need a track, a voice and visuals for each piece, consolidating three subscriptions into one credit pool is a real saving in both money and workflow friction.
Small marketing and e-commerce teams producing repetitive product creative benefit similarly, particularly where localisation into several languages is part of the job.
Independent musicians and hobbyists are served by the music stack, especially the stem and MIDI tools that assume you know what you are doing.
Developers who need audio generation as an API — music, covers, TTS, cloning, voice changing — can use it as a backend without touching the web interface.
Who should be cautious: anyone whose work has meaningful legal exposure. Agencies producing client work, anyone publishing commercially at scale, and anyone whose output could be mistaken for a real person's voice should read the licensing and rights section below very carefully before adopting this platform. Also cautious: buyers who want price predictability, given the discount-then-double structure and the narrow refund terms.
TopMediai is primarily a web platform requiring no installation — every tool runs in the browser. There are also mobile applications, with the site footer linking to both Google Play and the App Store. The privacy policy notes that the app may access your photo gallery or camera with your consent, to support AI generation features, and states that this access is required for core functionality.
For developers, the five audio APIs listed above provide programmatic access with published documentation.
On governing law, the terms take an unusually layered approach: they name the laws of the United Kingdom, the State of Colorado, and the applicable local laws of the user's jurisdiction in the Asia-Pacific region, with conflicts resolved by user location — Colorado law for Colorado users, local law for Asia-Pacific users, and UK law for everyone else. If you ever need to enforce anything, which of those applies depends on where you are.
Everything is credit-based, and credits are fungible across the whole platform. Prices below are from the official pricing page at the time of writing; several tiers were running promotions.
Starter Plan — billed weekly. First week 50% off at $4.99, $9.99 or $14.99 per week (regular $9.99, $19.99, $29.99), providing 1,500, 3,000 or 4,500 credits per week. Effective credit cost is $0.0033 in the first week, then $0.0067.
Creator Plan — billed monthly. First month 50% off at $8.99, $17.99 or $26.99 per month (regular $17.99, $35.99, $53.99), starting at 3,750 credits per month. Effective credit cost is $0.0024 in the first month, then $0.0048.
Pro Plan. $99.99, $149.99 or $199.99 (regular $199.99, $299.99, $399.99) at $0.0033 per credit.
There are also standalone tiers: a Free Trial, a Weekly Plan at $4.99, a Monthly Plan at $8.99 and an Annual Plan at $99.99 per year.
What credits actually buy. For the 1,500-credit weekly allowance, official conversions are: up to 150 songs, up to 83 videos, up to 300,000 characters of text to speech, up to 5 minutes of speech to speech, up to 15 voice clones, up to 50 song covers, up to 6 minutes of sync video, up to 50 minutes of audio enhancement and up to 30 minutes of video translation. Note how differently the tools consume — 150 songs against 5 minutes of speech-to-speech is a wide spread, so which tools you actually use changes the value of a plan enormously.
The discount-then-double structure. Almost every tier advertises a headline price that applies only to the first billing period, after which the rate roughly doubles. The Creator tier goes from $8.99 to $17.99; Starter from $4.99 to $9.99 per week. Your ongoing cost is the second number. This is corroborated as a durable structure rather than a one-off sale: the GoodFirms directory listing records the same pattern from an earlier snapshot, showing Starter at $4.99 renewing at $8.99 and Creator at $8.99 renewing at $16.99.
Weekly billing on the entry tier. This is the part of the credit pricing that catches people out: charging weekly is uncommon for creative software and multiplies the number of renewal events, which matters given the refund terms below.
The policy is headlined as a 30-day Money Back Guarantee, but reading the actual clauses, it is a conditional technical-fault refund rather than a no-questions-asked window.
Refunds are granted only in six circumstances, all of them about malfunctions or billing errors: the tool malfunctions due to an unresolvable technical problem within the trial period; the online tool suffers a fatal malfunction; the customer mistakenly duplicates a purchase; order information is not received within 24 hours of payment and support does not respond within 48 hours; the customer does not receive an auto-renewal notification before being renewed against their wishes, provided they make contact within 7 days; or duplicate payments occur due to a payment platform issue.
The ineligibility list is where the real constraint sits. Refunds are explicitly not available where "the customer purchased the product on special terms during a sales promotion." Since the pricing page shows nearly every tier running a 50% first-period promotion, this clause plausibly covers a large share of actual purchases. Also excluded: finding the product cheaper elsewhere afterwards, buying a similar product from another provider, "failing to properly understand the functions and capabilities of the product through not reading the product description thoroughly," and refusing to give a reason for the request.
Read together, the "30 days" refers to a trial window during which you are expected to find and report faults — not a period in which you may change your mind.
Because TopMediai spans three categories, the comparison set splits.
For AI music, the direct comparisons are Suno and Udio, which are more specialised and generally regarded as further ahead on song quality, but which do not give you video or voiceovers in the same balance.
For AI voice and TTS, ElevenLabs is the reference point for quality and for its more developed rights and consent tooling around voice cloning. PlayHT and Murf occupy adjacent positions. If voice is the bulk of your work, a specialist is worth pricing against this platform.
For AI video, you can go directly to the sources — Kling, Veo, Sora, PixVerse and the rest are the models TopMediai aggregates. Direct access may cost more or less depending on volume, but it removes an intermediary and gives you the vendor's own terms.
The honest summary: TopMediai's argument is not that any single capability is best in class, but that having all three behind one credit balance removes real friction and cost from a multi-tool workflow. If you genuinely need all three regularly, that argument is strong. If you need one of them seriously, a specialist will likely serve you better.
The licensing contradiction is the most important thing on this page. TopMediai's own product usage policy is admirably strict on paper. It states that Voice Clone "may not be used to clone any real person's voice without consent, including celebrities, influencers, public figures, family members, or private individuals," and designates voice cloning a high-risk feature subject to enhanced review. It states that AI Cover "may not be used for unauthorized covers of copyrighted songs, publication of celebrity voice covers without permission, misleading fake artist releases, or commercial use of cloned singer voices without proper rights," and that users are solely responsible for securing all necessary permissions.
Yet the homepage footer actively promotes and links to exactly that material. Under "Popular AI Song Covers" it lists Ronaldo AI Cover, Kim Tae Hyung BTS AI Cover, Max Verstappen AI Cover and Charli D'Amelio AI Cover — all real, identifiable public figures. Under "Popular Character Voices" it lists Mickey Mouse AI Voice, Darth Vader AI Voice, Sonic the Hedgehog AI Voice, Elsa AI Voice, Optimus Prime AI Voice and Bluey AI Voice — all copyrighted characters belonging to major rights holders.
The marketing surface and the compliance surface point in opposite directions, and the contract resolves the tension entirely at your expense. The terms state that you promise your input will not infringe anyone's intellectual property rights, portrait rights, reputation rights, name rights or privacy rights, and that if it does, the risks and liabilities are borne by you — along with the platform's economic losses, goodwill losses, rights-protection costs and attorney's fees. In plain terms: the site markets the feature, and you carry the legal risk of using it. Anyone publishing commercially should treat the celebrity and character pages as marketing rather than as permission.
The Trustpilot rating displayed on the homepage is no longer valid. The site shows "Topmediai.com Reviews 636, 4.6". Checking Trustpilot directly, the profile now carries a warning that the company's rating is unavailable due to a breach of guidelines, with the explicit statement: "We've removed a number of fake reviews for this company." The rating has been suspended. The profile still shows 654 reviews with a distribution of 87% five-star, 3% four-star, 1% three-star, 2% two-star and 7% one-star — but since the platform has determined that fake reviews were present and removed them, that distribution cannot be used to draw conclusions about quality. We are reporting the suspension rather than the number.
There are serious unproven allegations in the review record. One reviewer alleges that after leaving a one-star review, the merchant retaliated by manipulating their account credit balance and sent a phishing email from a fake Zendesk domain to obstruct a payment dispute. This is a single user's account, unverified, and we present it as an allegation rather than a finding. It is included because a prospective buyer would reasonably want to know it exists.
No coverage in major technology media. Searches found no reporting from TechCrunch, VentureBeat, The Verge, Ars Technica, ZDNet or comparable outlets. Nearly all "review" results are content-marketing pages published by direct competitors in the same category, which have a structural conflict of interest and were excluded. The practical consequence is that there is very little genuinely independent assessment of this product available.
International data transfer disclosure is thin. The privacy policy says only that information may be stored on secure servers outside your country of residence and may be transferred internationally, in compliance with applicable laws. It names no specific country or region where data is held and lists no sub-processors. Given that the uploads at stake include voice samples and likenesses — among the most sensitive categories of personal data — and that the operating entity is registered in Hong Kong with the Trustpilot business profile listing China as its location, that is a material gap for users in jurisdictions with data-residency expectations.
Nothing is stated about whether your uploads train the platform's models. Neither the privacy policy nor the terms address this in either direction. The closest language is processing your prompts to create generations and improving the services. If this matters to you, ask before uploading voice samples; we are not going to invent an assurance the company has not given.
The About page is significantly out of date. It still describes the company as "providing online ai video editing tools" and lists only text to speech, voice changer, voice cloning and vocal remover — omitting the AI music and AI video generation that now dominate the homepage. This is a small thing, but it suggests documentation is not maintained in step with the product.
Free-tier terms are deliberately indefinite. The official FAQ says free trials, credits, available models and export options "may vary by tool, plan, or usage type." You cannot determine your free allowance from the documentation; you have to discover it.
You may not use the outputs to build competing systems. The usage policy prohibits using platform outputs, models, voices or technologies to develop competing products, replicate platform capabilities or train external systems without written authorisation.
Corporate identity is not cleanly resolved. First-party terms say Cleverguard Technology Co., Limited; a third-party directory says Shenzhen iMyfone Technology Co. Ltd. The first-party source is more authoritative, but the divergence is unexplained.
Partially, and the terms are deliberately vague. The official FAQ states that TopMediai offers ways to try selected AI tools before upgrading, but that free trials, credits, available models and export options may vary by tool, plan or usage type. There is a Free Trial tier listed on the pricing page. In practice you should assume the free allowance is small and undefined, and measure your actual consumption before committing to a plan.
There is a single credit pool shared by every tool on the platform — the pricing page states that credits can be used across the entire platform. What differs is the exchange rate. For 1,500 credits, official "up to" figures are 150 songs, 83 videos, 300,000 characters of text to speech, 5 minutes of speech to speech, 15 voice clones, 50 song covers, 6 minutes of sync video, 50 minutes of audio enhancement and 30 minutes of video translation. Because those rates vary so widely, the value of a given plan depends heavily on which tools you actually use.
Because the advertised price is a first-period promotion. Nearly every tier runs a 50% discount for the first week or month and then reverts. Creator goes from $8.99 to $17.99 per month; Starter from $4.99 to $9.99 per week. Note also that the entry tier bills weekly rather than monthly. Budget using the post-promotion rate.
Probably not, despite the "30-day Money Back Guarantee" label. Refunds are granted only for technical faults, duplicate charges, undelivered orders, or auto-renewals made without prior notice (contacted within 7 days). Critically, the policy explicitly excludes purchases made "on special terms during a sales promotion" — and nearly all tiers are promotional. Not understanding the product's capabilities before buying is also listed as ineligible. Read this before purchasing rather than after.
The platform's own policy says no. Its product usage rules prohibit using Voice Clone on any real person without consent, explicitly including celebrities, influencers and public figures, and prohibit AI Cover from being used for unauthorised covers of copyrighted songs or publication of celebrity voice covers without permission. This is worth stressing because the homepage simultaneously promotes pages named after real celebrities and copyrighted characters. The terms place all infringement liability on you, including the platform's legal costs. Treat those pages as marketing, not as permission.
The terms and privacy policy both state that topmediai.com is operated by Cleverguard Technology Co., Limited, with offices in Hong Kong, London and Denver, each with a registration number. A third-party directory instead lists Shenzhen iMyfone Technology Co. Ltd, which we note without reconciling. On data location, the privacy policy says only that information may be stored on servers outside your country of residence and transferred internationally in compliance with applicable law — it names no specific country and lists no sub-processors.
The published documentation does not say, in either direction. Neither the privacy policy nor the terms address model training on user content; the closest statements concern processing your prompts to create generations and improving the services. If this is material to your work, contact the company directly before uploading voice samples or likenesses.
It functions largely as an aggregator on the video side. The published supported-model list includes Seedance 2.5, GPT-Image-2, Veo 3, Sora 2, Vidu 2.0, PixVerse, Wan 2.5, KlingAI, Nano Banana, Midjourney, Nano Banana Pro, Seedream, Grok and Hailuo. The audio stack — music generation, covers, TTS, cloning — appears to be the platform's own, which is consistent with the fact that its five public APIs are all audio APIs. Availability of any specific third-party model is a commercial arrangement and may change.
Text to speech reads a written script in a chosen voice from the library of 3,200+ voices across 190+ languages. Speech to speech takes a recording you have made and transfers your delivery — timing, emphasis, emotion — onto a different voice. If getting the prosody right matters, speech to speech usually gets you there faster than adjusting a TTS script. Note that speech to speech is one of the more credit-expensive tools relative to its output length.
No. Checking Trustpilot directly, the profile carries a warning that the company's rating is unavailable due to a breach of guidelines, with the platform stating that it has removed a number of fake reviews for this company. The rating has been suspended, while the homepage continues to display 636 reviews at 4.6. The profile currently shows 654 reviews, but because fake reviews were identified and removed, the star distribution should not be used to judge quality.
