MoboDrama Review 2026: an All-AI Claim, and How to Test It

Updated October 3, 2026 · Tested by the AI Drama Reviews editorial team · 34 min read · Our score: 8.2 / 10

DramaBoxEditor's pick — biggest short-drama library, lowest price to go unlimited 9.8Excellent

• Thousands of series, roughly 200 new titles every month
• Free episodes daily, plus ad-unlocks — no card needed to start
• Unlimited plans from about $5.99 a week, the lowest entry price we found
• Runs on Android, iOS and in a desktop browser

Go to DramaBoxAffiliate link · free to download · in-app purchases available

Affiliate link. We may earn a commission if you install through it, at no extra cost to you — it never changes our scores or placements. How we test · Disclaimer

What is MoboDrama?

MoboDrama is a vertical short-drama app for Android and iOS whose catalogue is sold as entirely AI-generated, and it is the smallest-scale service in our twenty-three-app ranking. It ships as "MoboDrama: AI Comic Dramas" on the US App Store and Google Play. What you get is 9:16 video made for a phone: episodes of roughly 60 to 120 seconds, seasons of 40 to 100 or more, free opening episodes on each series, and then a paywall you clear with coins or an SVIP subscription.

On who publishes it, we can report only what the stores carry. We looked for a company page, a named executive, a funding record and an audited title count on August 19, 2026, and again when we re-read the listings on October 3, 2026, and found none of them. That is the result of our search rather than a finding about who is behind the app.

The product name is the whole of MoboDrama's public AI disclosure. The word AI appears there and nowhere else in its material, and whether the catalogue is in fact machine-made is our own dated judgement rather than anything the company has documented. We score the app 8.2, thirteenth of twenty-three. The evidence behind the all-AI claim and the eight on-screen tells you can run yourself are further down this page, and the longer account of format, scale and positioning comes next.

How much does MoboDrama cost?

SVIP costs $14.99 to $29.99 a week or $149.99 a year, and coin packs cost $9.99 to $49.99. We read those figures on MoboDrama's US App Store and Google Play listings on October 3, 2026. Verify at checkout

Installing the app costs nothing, and every series opens with free episodes before the wall.

The number that would let you budget a season is the one nobody publishes: how many coins an episode unlocks. MoboDrama does not state that rate, and neither does any other app in our ranking, so coins buy episodes at a price you learn as you spend. That is why we score billing transparency 6.0 and why this review quotes no season total. The full pricing section lists what we looked for and what came back, and the axis scores show how the 6.0 feeds into the 8.2.

Is MoboDrama safe?

The risk with MoboDrama is billing, not malware. It is distributed through the US App Store and Google Play, and it behaved like an ordinary freemium media app throughout our testing on August 19, 2026. Our testing was installation and viewing on US Android and iOS accounts rather than a security analysis of the application, and we have no evidence of malicious behaviour of any kind.

What can cost you money is the coin model, and it does it quietly. Coins are bought in packs of $9.99 to $49.99 and then spent one episode at a time, and the rate - how many coins an episode unlocks - is not published by MoboDrama or by any other app in this category. The spending therefore accumulates one tap at a time against a price list you cannot read in advance, and you find out what a season cost after you have paid it. Coins are one-off purchases: they do not renew, and in practice they are not refundable.

The second billing risk is the subscription renewing after you have stopped watching. SVIP at $14.99 to $29.99 a week renews until somebody cancels it, and cancelling happens in the store rather than in the app, which is where most people in this category get caught. The four steps are below.

How do I cancel MoboDrama?

Cancel MoboDrama in your App Store or Google Play subscription settings. There is no cancellation inside the app, and deleting the app does not cancel anything, because the subscription belongs to the store rather than to the app on your phone.

1. iPhone and iPad: Settings, then your name at the top, then Subscriptions, then MoboDrama, then Cancel Subscription.
2. Android: Play Store, then your profile icon, then Payments and subscriptions, then Subscriptions, then MoboDrama, then Cancel subscription.
3. Check that it took, in the store rather than in the app. The entry should now show an expiry date rather than a renewal date. In this category it frequently still shows a renewal date after a cancellation that appeared to go through, and if it says renews, it will renew. Cancel again and check a second time.
4. Coins cannot be cancelled, because they are not a subscription. They are one-off purchases, they do not renew, and in practice they are not refundable. Whatever is left unspent stays unspent.

You keep access until the period you have already paid for ends, so cancelling on the day you subscribe costs you nothing. The walkthrough for every app in the category is in how to cancel a short drama subscription.

MoboDrama in detail: format, scale and positioning

MoboDrama is a vertical short-drama app that positions itself as a fully AI-generated catalogue, and it is the smallest-scale service in our twenty-three-app ranking. The format is the category standard: 9:16 video for a phone, episodes of roughly 60 to 120 seconds, seasons of 40 to 100 or more episodes, a hook inside the first three seconds and a cliffhanger at the end of nearly every episode.

What separates MoboDrama from the larger apps is not the format. It is the pitch. DramaBox and ShortMax sell breadth, ReelShort sells English-language originals with US casts, StoReel sells a creation tool. MoboDrama sells the production method itself: everything you are watching, it says, was made by machines.

That is an unusual thing to lead with, because it is the one attribute in this category that no buyer can verify and no third party measures. It is also, commercially, a smart position in August 2026, when the term "AI drama" carries enough search interest to be worth owning. Both of those things can be true at once, which is why this review devotes an entire section to how a viewer can test the claim rather than take it.

How small is "smallest scale"?

Small enough that we will not pretend otherwise. MoboDrama has no Sensor Tower revenue figure attached to it, no download milestone, no funding round and no executive on the record about its production line. Compare that with the apps immediately above it: FlickReels has a Sensor Tower growth figure of 3.75x quarter on quarter for Q1 2025, VibeShort reported 3.95 million downloads in a 30-day window in May 2026 with year-on-year growth of 346.58%, and StoReel raised $34 million with a stated target of 100 AI dramas a month (36Kr, March 26, 2026).

MoboDrama has none of that. Our assessment therefore rests almost entirely on installing the app, watching it, hitting the paywall and writing down what happened. That is a legitimate evidence base, and it is a narrower one than we use on most of the reviews on this site. Treat the 8.2 as a considered judgement with wider error bars than the scores above it.

Is MoboDrama actually AI-generated?

MoboDrama says its catalogue is entirely AI-generated. We can neither confirm nor refute that, and neither can anyone else, because no authoritative dataset shows what percentage of any app's catalogue is AI-produced. That gap is not specific to MoboDrama. It is one of the declared holes in the evidence for this entire category, and we list it openly rather than filling it with a guess.

What we can report is what watching the app looks like. In our August 2026 testing the material we sampled was consistent with generated video throughout: shots that end at roughly the same short length, faces that move between takes, backgrounds whose written text is not written text, and the particular flat-lit, over-smooth surface quality that current video models produce when nobody spends money grading the output. Nothing we watched read as filmed with a camera.

That is evidence in favour of the claim. It is not proof of it, for two reasons worth stating plainly. First, a catalogue can be 90% generated and be described as all-AI without anybody noticing the difference from the outside. Second, and more importantly, generated-looking is not the same as generated: heavy AI dubbing, AI upscaling and AI restoration applied to filmed footage can produce several of the same artefacts.

Context for why an all-AI catalogue is entirely plausible in 2026: the China Netcasting Services Association reported via Global Times that more than 95% of new Chinese microdramas in Q1 2026 were AI-generated, and DataEye counted roughly 221,900 AI titles published on Douyin in the first half of 2026, an average of about 1,200 a day. In April 2026 alone Douyin carried around 44,200 AI titles against 3,248 live-action ones. A catalogue of the size MoboDrama runs could be stocked entirely from that pipeline without difficulty.

So the claim is credible. Credible is not verified, and the difference between those two words is the whole of the next section.

How to test an "all-AI" claim yourself, in about ten minutes

You cannot audit MoboDrama's catalogue, but you can gather real evidence about it in the time it takes to watch four episodes, using eight on-screen tells that current video models cannot fully suppress. This protocol is the original contribution of this page. It is not a detector and it does not produce a verdict of the kind a laboratory would sign. It produces evidence, weighted, that you can act on.

Do it with a specific method: open four titles from different shelves, watch the first two episodes of each, and watch on the largest screen available to you. A phone at arm's length hides most of what follows. A laptop or monitor does not.

Tell 1: shot length against the 10-12 second ceiling

Count the seconds between cuts. Generated drama rarely holds a single shot beyond about 10 to 12 seconds, and that is a model constraint rather than a stylistic decision. ByteDance's Jimeng produces roughly 12 seconds per generation, or 30 with motion templates. A scene in this format is assembled from many short generations rather than rendered as one continuous take.

The counter-example matters as much as the rule: Kling runs to about three minutes, so a long take does not by itself disprove generation. What it means is that the producer either paid for a longer-context model or filmed it. If every shot in a series lands between three and eleven seconds, and the cuts fall at suspiciously regular intervals, you are almost certainly looking at assembled generations.

Tell 2: identity drift between cuts

Watch one character across a conversation and check whether the face is the same face after every cut. Identity drift is the defining failure mode of generated video: a nose narrows, an eye colour shifts, an age reads three years younger in the reverse angle. It is subtle by design, because producers use character reference images, locked character bibles and fixed seeds specifically to suppress it.

Suppressed is not eliminated. Check the secondary characters rather than the lead, because reference images are usually built for the people who carry the series and not for the colleague with four lines. Check jewellery, hairstyle, collar shape and body proportions as well as faces. We explain the underlying mechanism in why AI drama characters change between shots.

Tell 3: hands, and Tell 4: background text

Hands and written text are where generation still fails most reliably, so freeze on both. Fingers fuse, count wrong, bend at joints that do not exist, or hold a cup that changes shape between frames. Signage, phone screens, book covers and newspaper headlines resolve into pseudo-letters that look like language until you try to read them.

These two tells are strong because they are expensive to fix. A producer can reshoot a face with a reference image; correcting every hand and every sign in a 70-episode season is manual work at a scale the economics do not support. If background text is unreadable and nobody fixed it, nobody could afford to.

Tell 5: the same faces across unrelated titles

Open three series from different genres and look for the same face in all three. Models trained on similar data produce similar faces, so an all-AI catalogue tends to feel cast from a very small agency - the same jawline playing a billionaire, a werewolf and a Qing dynasty official.

This is the tell that scales best across a catalogue rather than a title, which makes it the most useful one for testing MoboDrama's specific claim. A mixed catalogue draws its live-action titles from real casting and its generated titles from a model, so the faces diverge. A genuinely all-AI catalogue converges. Convergence is not proof either, but a catalogue in which every romantic lead is plausibly the same person is telling you something about how it was made.

Tell 6: lip movement against the dialogue

Mute nothing and watch the mouth. If lip movement does not track the words, note it - and then note that this tell is the weakest of the eight. Out-of-sync lips appear in generated video, and they appear just as often in filmed drama that has been dubbed into English by a voice model, which is most of what the large apps carry.

Use it as a supporting signal, never as the deciding one. It tells you an AI system touched the audio or the mouth. It does not tell you whether a camera was involved.

Tell 7: physics, object permanence and screen direction

Watch what happens to props and to space. A glass on the table in one shot is gone in the next, a bag changes shoulder, rain falls at an angle that a body does not respond to, a character walking left to right emerges from the cut walking right to left. Screen-direction reversal in particular is something a trained editor would never allow and a shot-by-shot generation pipeline produces constantly.

Tell 8: AI dubbing is not AI generation

This is the tell that stops the other seven from misleading you. The most common form of AI in this category is not generated video at all - it is machine translation, synthetic voice and automated dubbing applied to filmed Chinese-language series. An app can be soaked in AI and carry almost no AI-generated footage.

So when you file your verdict, separate the two questions. "Did AI touch this?" is nearly always yes across the whole category. "Was this footage generated?" is the question MoboDrama's claim is actually about, and only tells 1 through 5 and 7 speak to it.

What this protocol proves, and what it does not

It produces evidence, not proof. Four titles are not a sample of a catalogue, artefacts can be edited away by anyone willing to pay, and the absence of artefacts in a well-funded title says more about the budget than the method. If every one of the eight tells fires across four series from different shelves, "this catalogue is generated" is a well-supported conclusion. If none of them fires, you have learned that the production was careful, not that a camera was present.

The tooling that would settle it exists and is not being used at viewer level. C2PA records an asset's origin and edit history in the file itself, and Google's Veo 3.1 attaches a SynthID watermark to its output. Neither is a rights registration and neither is a detector you can run on a streaming app. Until a platform in this category voluntarily exposes provenance data per title, the eight tells are the best a viewer has. Our longer version of this method lives in how to tell if a drama is AI-generated.

One last point of intellectual honesty: we applied this protocol to MoboDrama and the results were consistent with an all-AI catalogue. We are reporting that as our observation of the titles we watched, in August 2026, and not as a finding about the catalogue as a whole. If MoboDrama published a per-title provenance manifest tomorrow, this section would become unnecessary, and we would say so happily.

How does MoboDrama score, axis by axis?

MoboDrama scores 8.2 out of 10. Its weakest axis is billing transparency at 6.0, one point below catalogue and refresh at 7.0. Weights and protocol are on our method page.

One note on how the 8.2 was reached. An all-AI catalogue is not itself a scored axis, and deliberately so. Our axes describe what a viewer receives - what a season costs, how much there is to watch, how it looks, what runs it, and whether anyone will tell you the price. MoboDrama's production method is the most interesting thing about it and the least relevant to any of those. It earns the app a review, not points.

How much does MoboDrama cost to finish one season?

MoboDrama publishes SVIP at $14.99 to $29.99 a week and $149.99 a year, and coin packs at $9.99 to $49.99. We read those figures on its US App Store and Google Play listings on October 3, 2026. Verify at checkout

Correction, October 3, 2026. This review previously said MoboDrama published no US price, which is what our August 19, 2026 search found. The store listings carry one now, so we have rewritten this section and raised billing transparency to 6.0.

Billing transparency scores 6.0 on the strength of those listings, which is mid-table here: the same mark ShortMax carries, above GoodShort and NetShort at 5.0, below DramaBox at 7.0 and ReelShort at 8.6. It goes no higher because the listings price the subscription and the coin packs but not the episodes: not one of the twenty-three publishes how many coins an episode costs, so a reader still cannot work out what finishing a season costs before starting it. Two of the twenty-three apps in our ranking publish no coin-to-dollar rate, GoodShort and NetShort, and MoboDrama is no longer one of them.

What the category charges, for context only

These are category-wide figures and not MoboDrama's numbers. Coins across this category run $0.20 to $0.50 an episode, which puts a 70 to 80 episode season at $14 to $40. Weekly unlimited plans across the category run $5.99 to $29.99. DramaBox sits at the bottom of that range at about $5.99 a week and StoReel near the top at about $19.99 to $29.99.

Use those as a sanity check. MoboDrama's $14.99 to $29.99 a week sits in the upper half of that band, for the smallest catalogue in the ranking, while a week of the largest catalogue in the category costs about $5.99. That is the comparison to make before you subscribe.

How much can you watch on MoboDrama for free?

MoboDrama is free to install and opens series with free episodes, following the category convention of the first three to ten before a wall. The exact allowance is not documented, which is consistent with everything else about this app's disclosure.

For MoboDrama specifically, the free tier matters more than it does for a larger app, and for an unusual reason: it is enough to run the entire eight-tell protocol in the section above. Four titles, two episodes each, is eight episodes - inside the free allowance if you spread it across series rather than burning it on one. You can therefore test the central claim of this app without giving it any money at all, which is exactly what we recommend you do.

The wall itself is placed where every wall in this category is placed: at the moment the first episode block ends on a reveal. That is engineering, not coincidence. The question worth asking at that moment is not whether you want to keep watching - you will - but whether a week of SVIP at $14.99 to $29.99 is worth a catalogue this size.

What is actually in the MoboDrama catalogue?

MoboDrama has the smallest catalogue of the twenty-three apps we rank, and publishes no refresh rate, no title count and no release schedule. For comparison, DramaBox carries thousands of series and adds roughly 200 new titles a month; MoboDrama's shelves run out in an evening of browsing.

That creates a tension worth naming. An app claiming fully AI production is claiming the cheapest production method available - FlexTV's vice-president told MIT Technology Review on May 15, 2026 that shifting North American production to AI cut costs by 80 to 90% and compressed schedules from three or four months to under one, with teams of about ten people. If MoboDrama's catalogue is genuinely all-AI, the natural consequence is a lot of titles, quickly. It does not have a lot of titles. Either the pipeline is smaller than the positioning implies, or the company is spending its savings on something other than volume.

What volume production does to hit rate

The Chinese data is the only place where AI-drama output is measured at scale, and it is unflattering about what cheap production buys. DataEye counted roughly 221,900 AI dramas published on Douyin in the first half of 2026, and found that about 1,055 of them - 0.47% - passed 100 million views.

Read that chart as an economic model rather than a scoreboard. Cheap production has not made hits more common; it has made misses cheap, which changes what a catalogue is for. A studio that can produce a 60 to 80 episode AI microdrama for $3,000 to $14,000 (NOW News, December 15, 2025) does not need to be right about what audiences want. It needs to be able to be wrong two hundred times a month.

The former Meta executive quoted in April 2026 put the resulting category better than we can: not a substitute for conventional drama, but "a parallel category - one defined by speed, variation, and the ability to cater to almost any audience preference imaginable." That is precisely what a viewer encounters on an all-AI shelf. Enormous variation, very little curation, and a hit rate that the platform absorbs and you experience as a lot of abandoned first episodes.

For MoboDrama that model is only half in place. It has the variation and the lack of curation. It does not yet have the volume that makes the model work, which is the clearest structural argument against installing it before FlexTV or StoReel.

How good is the video, audio and dubbing on MoboDrama?

MoboDrama looks like competent 2026 generated video: clean, flat, slightly plastic, and stable for as long as a shot lasts. We scored production quality 7.6, which places it below FlexTV and StoReel and above the weakest experimental slates. Nothing we watched was broken. Nothing we watched would be mistaken for a camera either.

The artefacts are the category-standard ones, in roughly the order we noticed them: identity drift between cuts on secondary characters, hands that do not survive a freeze frame, background text that is not text, lip movement running slightly ahead of or behind the English dialogue, and the shot-length ceiling that gives every scene the same cut rhythm. None of it is a MoboDrama failing specifically. All of it is what current models do.

Audio is the weaker half. Synthetic voice in this category has improved to the point where individual lines are convincing and paragraphs are not: emphasis lands on the wrong word, emotional peaks flatten, and two characters in an argument perform at the same intensity. Music and sound effects do their job. What is missing is the thing a human performance supplies, which is a reason to believe anyone in the scene wants anything.

Set against that, the honest counterweight from the evidence base: Holywater's Bogdan Nesvit has said that engagement metrics for AI content are in some cases comparable to live-action, and director Qingge Gao said in August 2026 that most microdrama viewers "don't care." Our own quality score is a judgement about craft. It is not a prediction about whether you will finish the series, and the two come apart more often in this category than in any other.

Where does MoboDrama play?

MoboDrama is a phone app: Android and iOS, with no desktop browser player found during testing. That costs it against DramaBox and ShortMax, both of which play in a browser, and it costs it something specific on this page - the browser is the best artefact detector a viewer has.

If you intend to judge the production rather than simply consume it, watch on the largest panel you own. Hand deformation, pseudo-text and identity drift that a phone hides at arm's length are unmissable on a tablet. MoboDrama's lack of a browser player is therefore an obstacle to exactly the exercise this review is about.

How does MoboDrama compare with the alternatives?

The right comparison set for MoboDrama is not DramaBox or ReelShort - it is the three other AI-native drama services: FlexTV, MyMuse and StoReel. All four are asking the same question of a viewer, which is whether generated drama is worth paying for, and they answer it with very different amounts of evidence. Character.AI Series is AI-native as well, but its watch-then-chat format makes it a different proposition rather than a rival shelf.

The row that decides it is the evidence row. FlexTV is the only app in this category where the AI production claim rests on a named executive on the record - and even that is a vendor describing its own results, which is a caveat we repeat every time we cite it. StoReel has a funding round and a stated production target attached to its claim. MoboDrama has a marketing position. That is not an accusation of dishonesty; it is a statement about how much you should be willing to pay for an unchecked claim, which is less.

Against MyMuse, MoboDrama's 8.2 to 8.1 gap is not meaningful, and the two apps differ in character rather than quality. MyMuse swings harder - genuinely strange titles, quality varying widely between them. MoboDrama is flatter and more consistent. Choose MyMuse if you want to see what the technology does when nobody is being careful, and MoboDrama if you want to see what an ordinary all-AI shelf looks like. Read our FlexTV review and our StoReel review before either.

What does MoboDrama do well?

What does MoboDrama do badly?

Three of those weaknesses are worth more than a bullet, because they are the ones that would change a decision.

Leading with an unverifiable claim is a choice, and it has a cost. MoboDrama could label individual titles. It could publish a provenance manifest, attach C2PA data, or state which model produced which season. Any of those would convert the claim from marketing into information. Choosing the catalogue-level version instead means choosing the version that cannot be checked, and a reader is entitled to notice which version a company picked.

The subscription is priced; the rate that decides a season is not. The store listings give you SVIP at $14.99 to $29.99 a week and coin packs at $9.99 to $49.99, which is more than GoodShort or NetShort will tell you. What no app here publishes, MoboDrama included, is how many coins an episode costs, so the coin route is still a purchase made without a total. MoboDrama is also small enough that no independent listings exist to check the store figures against.

Our own evidence base for this page is thinner than for any other review on this site, and we would rather flag it than disguise it. There is no Sensor Tower revenue line for MoboDrama, no download figure, no funding round, no executive interview. Every other review here triangulates our testing against at least one independent measurement. This one does not, and the 8.2 should be read with that in mind. If you have used MoboDrama and our account does not match yours, tell us - on this page more than most, we want the correction.

Who should install MoboDrama, and who should not?

Install it if…

• You specifically want to see an unmixed AI catalogue rather than a mixed one, and you have already looked at FlexTV.
• You are interested in the technology as a subject, not only as entertainment - this is a good place to run the eight tells.
• Fantasy, xianxia, mythology and period drama are your shelves. They are where generation looks best and where MoboDrama is strongest.
• You are willing to test on the free tier first and then decide whether $14.99 a week is worth this catalogue.

Skip it if…

• You want a lot to watch. This is the smallest catalogue in our ranking, and the larger apps are not close.
• You want to know what a whole season costs before you start one. The subscription is priced; the coin rate behind it is not.
• You mainly watch contemporary romance or family drama. Those are the shelves where generated video looks worst.
• You want the strongest AI-native app rather than the loudest claim. That is FlexTV at 9.2, or StoReel at 9.1.
• You watch on a laptop or a television. There is no browser player.

How do you try MoboDrama without paying?

1. Install and turn off in-app purchases first. The coin packs run to $49.99, so this is an app where we would do that before opening it rather than after.
2. Spread the free allowance across four titles from four different shelves, two episodes each. That is the sample the eight-tell protocol needs.
3. Watch on the biggest screen you have. A tablet if there is no browser player. Artefacts that vanish on a phone are obvious at that size.
4. Run the tells in order and write down what fires. Shot length, identity drift, hands, background text, recurring faces, lip sync, physics. Then ask the eighth question: is the AI in the picture or only in the voice track?
5. At the paywall, read the number carefully before anything else. Check it against the store listing, which on October 3, 2026 carried SVIP at $14.99 to $29.99 a week and $149.99 a year.
6. Compare that number with $5.99, which is roughly what a week of unlimited viewing on the largest catalogue in the category costs. If MoboDrama is asking materially more for materially less, the decision makes itself.
7. If you do pay, screenshot the purchase sheet and the store receipt. A figure you did not record is much harder to dispute later.

Cancelling MoboDrama: the long version

Cancel in the store, not only inside the app. Switching a setting off inside a short-drama app frequently does not stop the billing, and that is the single most common billing mistake in this category.

1. iPhone and iPad: Settings, then your name at the top, then Subscriptions, then MoboDrama, then Cancel Subscription.
2. Android: Play Store, then your profile icon, then Payments and subscriptions, then Subscriptions, then MoboDrama, then Cancel subscription.
3. Confirm it worked. The store should show an expiry date rather than a renewal date. If it still says renews, it will renew.
4. If you bought coins rather than a subscription, there is nothing to cancel. Coins are one-off purchases, they do not renew, and in practice they are not refundable. That cuts both ways: no surprise renewals, no way back either.

You keep access until the period you already paid for ends, so cancelling on the day you subscribe costs you nothing. The full walkthrough for every app in the category is in how to cancel a short drama subscription.

Frequently asked questions

Is MoboDrama worth installing? The verdict

MoboDrama scores 8.2, and the claim it leads with is the most interesting thing about it and the least verified. A fully AI catalogue is entirely plausible in 2026 - more than 95% of new Chinese microdramas in Q1 2026 were AI-generated, and Douyin absorbed about 1,200 new AI titles a day in the first half of the year. Plausible is where the evidence stops. No dataset establishes the AI share of this app or any other, and MoboDrama has published nothing that would let anyone check.

What you can do is check it yourself, and that is the reason this page exists. Eight on-screen tells, four titles, two episodes each, on the largest screen you own, entirely inside the free tier. That produces evidence rather than proof, and evidence is what is actually available in this category. Anyone offering you a percentage is guessing.

As a place to watch, MoboDrama is a second stop rather than a first. It is the smallest catalogue in our ranking, it publishes no coins-per-episode rate and no refresh rate, and it has no independent record of any kind behind it. FlexTV at 9.2 gives you a better AI-produced slate with a production line that a named executive has described publicly. StoReel at 9.1 gives you a bigger one and a creation tool, at about $19.99 to $29.99 a week against MoboDrama's $14.99 to $29.99.

The recommendation is therefore specific rather than general. Install MoboDrama free, watch eight episodes, run the tells, and learn something about what generated drama can and cannot do. Then decide whether to pay - knowing what a week costs, knowing that nobody will tell you what a season on coins costs, and knowing that the app is asking you to take its central claim on trust.

DramaBoxEditor's pick — biggest short-drama library, lowest price to go unlimited 9.8Excellent

• Thousands of series, roughly 200 new titles every month
• Free episodes daily, plus ad-unlocks — no card needed to start
• Unlimited plans from about $5.99 a week, the lowest entry price we found
• Runs on Android, iOS and in a desktop browser

Go to DramaBoxAffiliate link · free to download · in-app purchases available

Affiliate link. We may earn a commission if you install through it, at no extra cost to you — it never changes our scores or placements. How we test · Disclaimer

Related reading

Sources

• Our own testing — installation on US Android and iOS accounts, free-tier exhaustion, the eight-tell protocol applied to four titles across four shelves, and a search for published US pricing, all on August 19, 2026. Protocol on how we test. This is the primary evidence base for this page, and it is thinner than for any other review on this site.
• MoboDrama — in-app store listing and app-level positioning, browsed August 19, 2026; US App Store and Google Play listings re-read October 3, 2026, which publish SVIP at $14.99 to $29.99 a week, $149.99 a year and coin packs at $9.99 to $49.99. MoboDrama does not publish an official US pricing page, a coins-per-episode rate, or labels marking individual titles as AI-generated. Its all-AI catalogue description is a company claim.
• China Netcasting Services Association, via Global Times (2026) — more than 95% of new Chinese microdramas in Q1 2026 were AI-generated. Chinese market data; the only market where AI-drama output is measured at scale.
• DataEye (2026) — approximately 221,900 AI dramas published on Douyin in H1 2026, roughly 1,200 a day, of which about 1,055 (0.47%) passed 100 million views; April 2026 on Douyin, ~44,200 AI titles against 3,248 live-action.
• MIT Technology Review, May 15, 2026 — Tang Tang, vice-president of FlexTV, on moving North American production to AI: costs down 80–90%, schedules from three or four months to under one, teams of about ten people. This is a vendor describing its own results.
• 36Kr, March 26, 2026 — StoReel's $34 million raise and its stated target of 100 AI dramas per month. Used here only as the comparison point for MoboDrama's undocumented scale.
• NOW News, December 15, 2025 — an AI microdrama of 60–80 episodes produced for $3,000–$14,000, the figure behind the volume economics described above.
• ByteDance Jimeng and Kling model documentation, 2026 — generation length limits of roughly 12 seconds per generation (30 with motion templates) and about 3 minutes respectively. These are the numbers behind the shot-length tell.
• Google, Veo 3.1 documentation, 2026 — SynthID watermarking on generated output; and the C2PA provenance standard, which records an asset's origin and edits and is neither a rights registration nor a detector. Both are cited to explain why viewer-level verification is currently manual.
• Vertical Drama Love, reviewer Jenny Cooper, June 24, 2026 — 25 AI titles watched, three finished, with fantasy identified as the strongest category. Cited as one reviewer's stated experience, not as survey data.
• PetaPixel and Today, August 2026 — director Qingge Gao on microdrama viewers being largely indifferent to AI production; and TheWrap, 2026 — Holywater's Bogdan Nesvit on engagement metrics for AI content being in some cases comparable to live-action. Both are expert opinion, not measurement.
• ACM, A Survey: Spatiotemporal Consistency in Video Generation (May 18, 2026) and OpenReview on narrative and temporal consistency in long video (March 26, 2026) — the academic basis for the identity-drift and continuity failure modes described in the tells.

Affiliate disclosure. We earn nothing from MoboDrama. There is no affiliate relationship, no commercial arrangement and no contact of any kind between us and this app, and the DramaBox links elsewhere on this page are the ones we are paid for. That arrangement is worth stating precisely here, because it cuts against us in an obvious way: this review recommends running MoboDrama's free tier and then comparing its $14.99 a week against DramaBox at about $5.99 a week, and DramaBox is the app that pays us. We think the comparison is correct on the numbers, and you should weigh it knowing who benefits from it. Accuracy notice. MoboDrama publishes SVIP and coin-pack prices in its US store listings, but no coins-per-episode rate and no title count. Free-episode allowances, catalogues and pricing in this category change without notice and vary by region, device, store and promotion. Pricing here was verified on October 3, 2026 and everything else on August 19, 2026, and the price the app quotes you at checkout is the only binding one. On the all-AI claim. We report it as a company claim throughout and never as an established fact, because no independent dataset measures the AI share of any catalogue in this category. Independence. We are not affiliated with MoboDrama and this is not an official support channel — for account or billing problems, contact MoboDrama directly. Editorial opinion. All reviews, scores and rankings on this site represent the editorial opinion of the AI Drama Reviews team, based on publicly available information and our own testing. They are not endorsements, guarantees of quality or professional recommendations. We do not operate or control any listed application and bear no responsibility for its service, billing or content.