Every video platform grew captions. The phone never did. Three working setups — same device, second device, or a captioned-telephone service — and how to pick.
Zoom has captions. Teams has captions. Even Instagram stories have captions. But the medium where mishearing costs the most — the phone call with the doctor's office, the insurance adjuster, the client who only does voice — still ships exactly the audio it shipped in 1995. If you need captions on a phone call, the phone will not provide them. Something listening to the call has to. The good news: that is a solved problem, and you likely have everything it requires already.
Captioning a video meeting is easy because the platform owns the audio: it can fork the stream to a transcription engine before it ever reaches your speaker. A phone call has no platform in that sense — just two devices and a carrier in between. Nothing in the path is offering to transcribe.
So every real solution does the same thing: it gets the call audio to a transcription engine some other way. There are exactly three ways to do that, and they differ in setup effort, caption quality, and who can use them.
This is the cleanest setup, and it covers more calls than people expect — most "phone calls" in working life now terminate on a computer anyway: softphones, VoIP desk lines, Google Voice, WhatsApp and Telegram desktop calls, and every call a scheduling link turns into.
Any app that plays the caller through your speakers works: a softphone, a VoIP client, a messenger's desktop app, or a phone-system web console. If your desk phone can forward to a softphone, forward it.
In Unicaption, pick your audio source: system or tab audio captures the caller's side directly — no microphone in the loop, so keyboard noise and room echo never touch the transcript.
Set the language you expect to hear, or leave detection on. If you want the captions translated live, set a target language and read both side by side.
Captions appear moments behind the speaker. If the call has domain vocabulary — drug names, case numbers, product SKUs — add them to the custom dictionary before dialing, and the engine will stop guessing.
When the call genuinely lives on a cell phone or a desk handset, use the two-device pattern: put the call on speaker, and let a laptop, tablet, or second phone run the captioning tool through its microphone.
It is less elegant than system audio and more sensitive to conditions, so stack the conditions in your favor:
This is also the only pattern that works for in-person-plus-phone situations — a family member on speaker at the kitchen table, a conference room with a caller on the Polycom — because the microphone hears the room and the call together.
In the United States, IP Captioned Telephone Service (IP CTS) provides captioned calls funded through the FCC's Telecommunications Relay Service program. Providers offer apps and dedicated phones that show captions during calls at no cost to the user.
The catch is eligibility and scope: the program exists for people with hearing loss and requires self-certification, the captions are typically English or Spanish, and it only covers phone calls — not meetings, not in-person conversations, not other languages. If you qualify, use it; it is a good program. If your need is broader — multilingual calls, translation, professional interpreting work, or captioning everything else in your day — a general-purpose tool covers the phone call and the rest.
The setups above change very little when the call is bilingual — which is where captioning a phone call stops being an accessibility feature and becomes an interpreting one. An interpreter working a medical or legal call by phone has no video, no lips, no slides: audio is the entire signal. A live transcript with translation beside it is the difference between working from memory and working from text.
For medical calls specifically, the compliance questions to ask of any tool in the audio path are covered in our HIPAA captioning guide.
| Your situation | Best setup | Why |
|---|---|---|
| Calls reach your computer (VoIP, softphone, messenger) | Same device, system audio | Cleanest audio, best captions, nothing extra on the desk |
| Calls live on a cell phone or handset | Speaker + second device | Works with any phone ever made; quality depends on the room |
| US, hearing loss, mostly English calls | IP CTS | Purpose-built and free — use it, and add a general tool for everything it doesn't cover |
| Bilingual or interpreted calls | Same device + live translation | Original and translation side by side; nothing stored |
Yes. Put the call on speaker and run a captioning tool on a second device — a laptop or tablet next to the phone. Keep the devices close and the room quiet. If the call can move to your computer (VoIP, WhatsApp desktop, a softphone), captions get noticeably better because the tool can read the call audio directly instead of listening through a microphone.
Live captioning with Unicaption adds nothing to the call itself — no bot joins, no announcement plays. Whether you should tell the other party is a separate question: consent norms and recording laws vary by state and country, and captioning that stores nothing is not recording. When in doubt, say what you're doing; it takes one sentence.
Yes. Unicaption shows the original transcription and a live translation side by side, in 60+ languages, with automatic language detection for callers who switch mid-sentence. Captions run moments behind the speaker.
With Unicaption, yes: sessions are end-to-end encrypted, audio is never stored, and transcripts are deleted when the session ends. Nothing about the call persists. That is the standard to hold any captioning tool to before it touches a sensitive call.
Live captions and translation for any call your computer can hear. 30 free minutes every week — no credit card.
Start Free Trial →