How to Choose a Closed Captioning Service

A closed captioning service turns the spoken words and relevant sounds in your video or audio into timed text that viewers can switch on or off. Providers differ in how they produce captions, how quickly they deliver them, which file types they export and how they treat your recordings. Settle what you need first, then compare providers against that list rather than against their marketing claims.

Start with the intended use

Captions for a recorded training video and captions for a live webinar are different jobs. Some providers handle only one. Before you request quotes, write down where the content will appear, whether it is recorded or live, and what you need to receive at the end.

• Is the content prerecorded, streamed live, or a live video call?
• Where will viewers watch it: a website player, a video platform, a meeting app, a broadcast channel or an internal learning system?
• Do you need a caption file, captions burned into the video, a transcript, or several of these?
• Which languages are spoken, and do you also need translated subtitles?
• How many hours of content do you expect each month, and how quickly must captions be ready?
• Do you have to meet an accessibility policy, a contract requirement or a regulation that sets caption quality standards?
• Does the material contain confidential, personal or unreleased information?

Source audio matters regardless of the provider. Clear speech recorded with a close microphone, little background noise and few people talking over each other is easier to caption accurately, whether a person or software does the work. If your recordings come from a laptop microphone in an echoing room, expect more corrections, and tell providers about it when you request a sample.

Features that matter

Human, automatic or hybrid captioning

Automatic speech recognition is fast and usually cheaper, but it may struggle with accents, technical terms, names, crosstalk and poor audio. Human captioners can interpret context and fix those problems, though they take longer. Many services combine the two: software produces a draft and a person reviews it. Ask which method each service tier uses, because the same provider may offer several tiers with very different results.

Accuracy claims

An advertised accuracy percentage is a provider statement, not a measurement of your content. Ask how it is calculated, what kind of audio it was measured on, and whether it counts punctuation, speaker labels and sound descriptions. Check whether the service guarantees a level of accuracy, and what it does if a file falls short: a free correction, a refund or nothing.

Caption content beyond the words

Proper closed captions do more than transcribe speech. They identify speakers when it is not obvious who is talking, describe meaningful sounds such as music, laughter or an alarm, and are timed and split into lines viewers can read comfortably. Ask whether these elements are included by default or cost extra, and whether you can supply a glossary of names and terms in advance.

Turnaround and live delivery

For recorded content, compare standard and rushed turnaround options and whether the deadline is counted in business days or calendar days. Live captioning is a separate service. It may be provided by human real-time captioners, often called CART, or by automatic live captions. Ask about the expected delay between speech and on-screen text, how the captioner receives the audio, what happens if the connection drops, and whether you get a cleaned-up file after the event.

Editing, languages and file handling

A built-in editor lets you correct a name or adjust timing without reordering the whole job. If you need other languages, ask whether translations are done by people, machines or both, and whether they are produced from the reviewed captions. Treat privacy as something to verify in writing: who can access your files, whether subcontractors or freelancers handle them, how long recordings and transcripts are kept, whether they are used to train speech models, and how you can request deletion. If the material is sensitive, ask whether the provider will sign a confidentiality agreement.

Check compatibility

Confirm the caption formats your destination accepts before you choose a service, then make sure the provider exports them. Common formats include:

• SRT: a simple timed-text format accepted by many video players and platforms.
• WebVTT: designed for web video players and supports some positioning and styling.
• SCC: associated with broadcast-style captions and often requested for television delivery.
• DFXP or TTML: XML-based formats used by some streaming and broadcast workflows.
• Plain text or document transcripts: useful for searchable records, but they are not caption files.

Next, check how files get in and out. Some services connect directly to video hosting platforms, cloud storage or meeting apps, so captions can be ordered and returned without manual uploads. Ask which integrations exist today rather than which are planned. Confirm the media formats and maximum file sizes the service accepts for upload. For live meetings, check whether captions appear inside your meeting app, in a separate browser window or through a streaming encoder, and test that route on the operating systems and devices your audience uses. If you burn captions into the video instead of supplying a separate file, viewers cannot switch them off or change their appearance, so keep a separate caption file as well.

Build a shortlist

Turn your requirements into a checklist and send the same questions to each provider. Any service that cannot give a clear answer on a must-have item should drop off the list.

• Supports your content type: recorded, live, or both.
• Offers the captioning method you need: human, automatic or reviewed automatic.
• Explains how its accuracy figure is defined and what happens when a file falls short.
• Includes speaker identification, sound descriptions and readable line breaks.
• Meets your turnaround needs at your expected monthly volume.
• Exports the caption formats your platforms accept.
• Integrates with your video host, storage or meeting app, or provides a workable upload process.
• Covers the languages and translations you need.
• Provides an editor or a correction process.
• Documents retention, access, subcontracting, model training use and deletion in writing.
• Pricing is clear about per-minute rates, minimums, rush fees and charges for extra features.

Finally, send each shortlisted provider the same short, representative clip, ideally one with the accents, terminology and audio quality you deal with every day. Compare the returned files in your actual player: check names, punctuation, timing, speaker labels and sound descriptions, and note how long the corrections took. That comparison, along with the written answers to your checklist, is a more reliable basis for a decision than a provider's own accuracy claims.