3Play Media makes every word count.

When Yale came calling, the founders of 3Play Media had a headquarters problem. Their office was an apartment in Somerville, Massachusetts. They lit candles before the visitors arrived. It was an improvised setting for a serious proposition: making online lectures accessible without turning every hour of video into an exhausting manual job. Yale became an early major customer outside MIT.

A useful machine still needs a correction

The venture grew from work with MIT OpenCourseWare. Four Sloan students - Chris Antunes, Josh Miller, C. J. Johnson, and Jeremy Barron - developed their captioning system and launched in 2008. Their insight was practical: speech recognition could produce a draft, while a better editing interface could reduce the work required to fix it. Accuracy and throughput belonged in the same design problem.

“It’s about creating accurate captioning at scale.”

C. J. Johnson / MIT News, 2015

Years later, Miller described the unusual advantage in an interview: the team had a customer before a product. That order matters. An institution already had lectures and a reason to make them usable. The founders could work against a concrete publishing problem. The resulting business still sells a way through that problem, now across far more kinds of content.

The operation has grown beyond its campus beginnings. Riverside Partners took a majority stake in 2019, when the investment announcement described more than 2,500 customers. In 2022, 3Play acquired Captionmax and National Captioning Canada, broadening its North American footprint and live-captioning capabilities. Financial terms were undisclosed. Today, 3Play reports more than 10,000 customers. That expansion helps explain the platform’s breadth: it serves organizations with very different content, audiences, and deadlines.

The video has more than one door

Consider a recorded lecture. A deaf viewer may need captions, including meaningful sounds and speaker identification. A blind viewer may need audio description explaining important visual action. A viewer who speaks another language may need translated subtitles or a dubbed soundtrack. A student revising for an exam may simply want to search the transcript and jump to the relevant moment.

3Play supplies those services through an online platform, integrations, and APIs. Its position sits between automated speech tools, captioning providers, and localization studios. Rev and Verbit are named alternatives in its own FAQ. The purchasing question is how much review, coordination, and delivery a team needs alongside the words. A university library and a single social clip create different jobs.

Sometimes the first failure is the org chart

At the University of Pennsylvania, different schools and centers were already using 3Play independently. The company’s case study describes fragmented processes and uncertainty about where to get support. Penn’s Information Systems and Computing department brought the service under one main account, created separate projects for participating units, and used a Panopto integration to automate captioning work.

The repair required setup, training, and collaboration. Buying software alone did not supply those things. FutureLearn offers another revealing example: one project used bulk transcription for 300 videos and converted 700 caption files from SRT to VTT. Some of the value was new text. Some was getting existing text into the required format. Video operations can be stubbornly literal.

Buy the review your video needs

3Play generally charges by the minute of uploaded content. Quotes depend on service, language, volume, turnaround, and the blend of AI and human review. Express targets occasional projects, Pro serves recurring smaller needs, and Enterprise adds larger workflows and customization. The bill follows both the content and the attention it requires.

Dubbing makes that tradeoff visible. The product page advertises four starting rates: Launch at $3 per minute, Refine at $7, Creator at $18, and Studio at $35. Each adds a different depth of human review. At those starting rates, ten minutes means $30 through $350. A project quote still matters; these are entry prices, observed in October 2026.

The backlog gets its own product

In March 2026, 3Play introduced Pulse. The target is the accumulated video library: content scattered across departments and platforms, with uneven captions and uncertain audio-description needs. Pulse scores caption accuracy, assesses description requirements, and routes files needing attention to human remediation. Its promise is a more selective review queue. That can help a team decide where to spend its next editing hour.

The approach also clarifies a limit. An automated caption track does not, by its presence alone, establish that a video is accessible. Nor does a score remove the need for institutional oversight. Teams still need to understand their content, review requirements, and publishing environment. Pulse gives them another operating tool.

First find the audience. Then find the voice.

Localization extends the same logic. 3Play announced human-reviewed AI dubbing in April 2024, then a YouTube creator offering in April 2026. The newer service combines dubbing with channel analytics to guide video and language selection. The useful shift is from translating everything to choosing what deserves the investment. Human involvement can cover scripting, cultural adaptation, and the delivery of the voice.

For a team considering the service, the copyable lesson is straightforward: connect one publishing workflow, test representative content, and judge the finished result with its intended audience. A short-lived clip may warrant a lighter tier. Dialogue whose humor or performance carries the whole video may warrant deeper review. Clean audio, terminology, and a clear quality target give that decision something solid to stand on.

3Play’s bet is that the work after the first draft remains valuable. The viewer encounters only the finished caption or voice. Getting that small thing right can require a surprisingly large operation.