Audio Drama: What It Is, Why It Is Back, and How to Make One

An audio drama is a scripted fiction story told entirely through voice, sound effects, and music, with no visuals at all. Characters are performed by voice actors, scenes are built from sound design rather than sets, and the listener assembles the picture. The format covers everything from a six-part supernatural mystery to a serialized comedy to a branded story a museum hands out on a walking tour.

Producing one traditionally meant a cast, a booth, and an editor. Tools like Jellypod change that arithmetic: you write the script, assign a distinct AI voice to each character, layer ambience underneath, and publish the finished series to Spotify and Apple Podcasts through hosted RSS. The format is also getting a commercial tailwind from an unexpected direction. Deloitte projects in-app micro-series revenue will more than double to about $7.8 billion in 2026, which means serialized short fiction is now a proven business, not a passion project.

What is an audio drama?

An audio drama is fiction performed for the ear. A script assigns lines to named characters, voice actors perform them, and sound design carries everything a camera would otherwise show: a door, a storm, a room's size, the distance between two people. There is no narrator explaining what happened unless the story deliberately uses one.

The form is old. BBC Radio Drama has produced scripted audio since the 1920s and remains one of the largest commissioners of new drama in the world, most of it on Radio 4. What changed is distribution. Podcast feeds removed the need for a broadcast slot, so a two-person team can now reach the same listener a national network reaches, in the same app, on the same commute.

Genre is wide open. Horror and science fiction dominate the best-known shows because sound is unusually good at suggesting things you cannot see, but comedy, romance, historical fiction, and documentary-style fiction all work. The constraint is not subject matter. It is that every plot point has to arrive through something a listener can hear.

What is the difference between an audio drama and a regular podcast?

A regular podcast is people talking as themselves. An audio drama is people performing as someone else. That single difference cascades into almost every production decision.

A conversational podcast is recorded once and lightly edited, and its appeal is the host's real personality. An audio drama is written first, then performed, then built in post, where sound effects and music do structural work rather than decorative work. A chat show can survive a flat episode because listeners came for the hosts. Fiction cannot, because listeners came for the story and a flat episode breaks it.

The other difference is serialization. Interview podcasts are usually episodic, meaning any episode works as an entry point. Audio dramas are usually serialized, meaning episode seven assumes you heard episode six. That makes the ordering of episodes part of the product rather than an accident of publishing date, which is worth planning before you release anything.

Are audio dramas still popular in 2026?

The audience underneath them keeps growing. Edison Research's Infinite Dial data, compiled by Pew Research Center, found 42% of Americans aged 12 and older had listened to a podcast in the past month as of 2023, up from 12% a decade earlier. Weekly listening rose from 7% to 31% over the same period. Podcast fiction sits inside that audience, distributed by the same apps and discovered through the same charts.

Prestige interest is real too. Homecoming, a Gimlet audio drama, was adapted into an Amazon series starring Julia Roberts. Limetown became a Facebook Watch series. Adaptation is not a reliable business plan, but it does establish that scripted audio can originate intellectual property rather than only reformat it.

The stronger signal in 2026 is coming from video, not audio, and it is worth understanding because the underlying listener behavior is the same.

Why is serialized short-form fiction booming right now?

The clearest evidence that people will pay for serialized fiction on a phone comes from micro-dramas, the vertical video format that exploded first in China and then globally. Sensor Tower's State of Short Drama Apps 2026 put the category at roughly $750 million in in-app purchase revenue in Q1 2026 alone, up 20% year on year, against more than 850 million downloads, up 140%. Average daily time in these apps reached 25 minutes globally by April 2026.

Deloitte's 2026 technology and media predictions forecast that in-app micro-series revenue will more than double to about $7.8 billion this year. ReelShort, the category leader in Western markets, passed $1 billion in cumulative revenue and turned its first profit in 2026. The money is no longer speculative.

Traditional media noticed. Fox Entertainment took an equity stake in the Ukrainian studio Holywater and committed to producing more than 200 vertical video titles, and Holywater went on to raise a $22 million Series A in January 2026, the largest microdrama investment outside Asia. AI is doing much of the production: MIT Technology Review documented how Chinese short dramas became AI content machines, with generated video replacing conventional shoots at scale.

None of that proves audio drama will follow the same curve. It does prove the thing audio drama creators have always been told is niche, serialized scripted fiction consumed in short sittings on a phone, is one of the largest growth categories in mobile entertainment. The format was never the problem. Distribution and production cost were.

Didn't Quibi prove short-form storytelling does not work?

Quibi is the obvious objection, and the comparison is worth making because the differences explain why one model collapsed and the other is compounding.

Quibi raised $1.75 billion, launched in April 2020, and announced its shutdown six months later, going dark that December. It sold a $4.99 monthly subscription (or $7.99 without ads) for episodes of 10 minutes or less, produced at prestige-television budgets with prestige-television talent, watchable only on a phone and, for most of its life, impossible to clip or share.

Micro-dramas inverted almost every one of those choices. They cost a fraction as much to produce, they sell single episodes for cents instead of a monthly subscription, they engineer a cliffhanger into the end of nearly every episode rather than trusting star power to bring viewers back, and they buy their audience through performance marketing instead of waiting for a brand to pull one in. Quibi asked people to subscribe to a network. Micro-dramas ask people to pay fifty cents to find out what happens next.

The pandemic gets most of the blame for Quibi, and it deserves some, since launching a commute-shaped product into lockdown was terrible timing. The more useful lesson for anyone making serialized short fiction is about cost structure: Quibi paid film prices for content that had to compete with free. The format was never the problem.

How long should an audio drama episode be?

There are two working answers, and they come from different traditions.

Classic podcast audio drama runs 20 to 40 minutes per episode across seasons of six to twelve episodes, which suits a commute and matches how narrative podcasts are packaged. That length gives a scene room to breathe and a season room for an arc.

The micro-drama tradition runs the opposite way: 70 to 100 episodes of one to two minutes each, with a cliffhanger engineered near the end of nearly every episode, a structure CNBC described as flooding the market with cheap bets and letting audiences pick the winners. The total runtime is similar to a feature film. The difference is that it is cut into pieces small enough to fit any gap in a listener's day.

Pick based on how your audience listens rather than on what feels literary. A show aimed at a commute can hold 30 minutes. A show competing with a feed during a five-minute wait cannot, and short serialized episodes with hard endings will outperform a long one. If you are unsure, the shorter structure is the safer test, because you learn whether people come back after episode two instead of after episode one of eight.

How do you make an audio drama?

The production chain is the same whether you record a cast or generate the voices.

1. Write the script

   Assign every line to a named character and mark sound cues explicitly. A script that reads well silently can still fail out loud, so read it aloud before you commit to it.
2. Cast the voices

   Every character needs a voice a listener can tell apart with their eyes closed. Distinctness matters more than range, because listeners identify characters by timbre before they identify them by name.
3. Record or generate the performance

   Record actors in a booth or asynchronously, or generate the performance from the script. Either way, keep takes organized by character and scene, because reassembly is where projects stall.
4. Build the sound design

   Ambience establishes where a scene is happening. Effects mark what happens in it. Music signals how to feel about it. Most first-time audio dramas under-build ambience and over-build music.
5. Mix and publish

   Balance dialogue against effects so speech stays intelligible on phone speakers and cheap earbuds, then publish as a serialized feed where episode order is preserved.

The step that sinks most independent productions is not writing. It is coordinating a cast across schedules and then stitching separately recorded takes into something that sounds like people in the same room. That coordination cost is why so many promising scripts never reach episode three.

Can you use AI to make an audio drama?

Yes, for most of the chain, with one honest limitation.

AI handles casting and performance generation well. In Jellypod you can assign a distinct voice to each speaker, direct delivery with audio prompting for reactions like laughter or hesitation, layer ambience beds under a scene, drop in audio you recorded yourself, and adjust timing on a timeline before rendering. Hosting, RSS, and distribution to Spotify, Apple Podcasts, and YouTube are part of the same workflow, so a finished series does not need a separate publishing stack. If you want a vertical video version for social feeds, Shorts turns an episode into something you can post.

The limitation is direction. A generated performance is very good at clear, natural delivery and still weaker than a skilled actor at a specific dramatic choice: the pause before a lie, the line read that means the opposite of its words. For a lot of shows that gap does not matter, and for a few it is the whole point. Write the scenes that depend on that kind of subtlety knowing you may want a human take for those, and generate everything else.

Character consistency is the other thing to plan for. Serialized fiction only works if a character sounds the same in episode nine as in episode one, so lock voices at the start of a series rather than choosing them episode by episode.

How do audio dramas make money?

Four models, in rough order of how reliably independent shows use them.

Listener support through Patreon or a membership tier is the most common, usually offering early episodes, bonus scenes, or a behind-the-scenes feed. Advertising works once download numbers justify a rate, which typically requires consistent output over a long stretch. Licensing and adaptation is the outlier outcome that funded the best-known shows but should not be the plan. Branded and commissioned work, where a museum, a nonprofit, or a company pays for a scripted series as an audience-facing asset, is the most underrated of the four and the one where production cost matters most.

The micro-drama market added a fifth: per-episode unlocks. Video apps give away the first ten to twenty episodes and sell coins to unlock the rest, monetizing roughly 2% to 5% of viewers who then spend heavily. That mechanic depends on app-store infrastructure that podcast feeds do not have, so it is not directly portable to a podcast, but it is a live demonstration that serialized fiction converts to revenue when episodes end at the right moment.

A note on where this fits

Scripted audio is not only for entertainment. The same technique, characters, scenes, and a story rather than a lecture, is how a lot of instructional content gets people to finish it. Teachers use dramatized historical scenes, compliance teams use scenario-based dialogue instead of policy recitation, and health organizations use patient stories rather than pamphlet text. If that is your goal rather than fiction for its own sake, the education use case covers the workflow for turning existing material into episodes.

The short version

An audio drama is scripted fiction performed for the ear, and the reason to make one now is that serialized short-form fiction has stopped being niche: Deloitte puts in-app micro-series revenue on track to roughly double to $7.8 billion in 2026. The two things that always made the format expensive, coordinating a cast and paying for production, are the two things that got cheap.