Speech to text app for iPhone: a buyer's guide

A speech to text app for iPhone turns spoken words into typed text. It comes as Apple's built-in dictation, a dedicated voice keyboard, or a notes app you record into and copy from. The right pick depends on where you want the text to land, whether you want it cleaned up, and whether you want a history of it.

By Eric Friedman. Last updated 2026-09-08.

What people mean by a speech to text app for iPhone

Type "speech to text app iphone" into the App Store and three different kinds of product answer. The first is not really an app at all: it is Apple's own dictation, built into the system keyboard on every iPhone, with no download needed. Tap the microphone key wherever you can type, speak, and the words appear as you go.

The second is a dedicated voice keyboard, the closest thing to a voice to text app for iPhone that works everywhere you type. It is a third-party replacement or add-on for the system keyboard, and it records, transcribes, and often cleans up your words. The finished text lands in whatever app you were using, whether that is Messages, Mail, or a notes app of your own. The third shape is a standalone notes app with a record button. You speak into it, and it transcribes on its own screen. You then copy the finished text into an email, a message, or a document by hand, one dictation at a time.

Each shape answers a different question. System dictation answers how to avoid typing a text message by hand, with nothing to install beyond a toggle in Settings. A voice keyboard answers how to dictate cleanly inside any app you already use, often adding cleanup, a custom dictionary, and a history the system keyboard does not keep. A notes app answers how to capture a thought right now and shape it into something later. That fits when you draft long pieces and do not mind an extra copy-and-paste step. None of the three is wrong; each trades convenience for capability differently. The rest of this guide walks through what to check on any of them. Then it compares six named products side by side.

How to pick one in four questions

1. Which apps do you need it in? If the answer is every app you type in, you want system dictation or a keyboard extension. A separate notes app you copy from will not cover that.
2. Do you want the words cleaned up? Apple's dictation types exactly what you say, filler words included. A cleanup pass removes false starts and repeated words before the text lands.
3. Does processing location matter to you? Some apps process speech on your device by default. Others send audio to a server, sometimes with a setting to keep it local instead.
4. Do you want a record of what you said? A notes app or a keyboard app built around a feed keeps a searchable history. System dictation keeps none at all.

The ranked picks for iPhone, compared on processing, price, and keyboard support, live in the best voice keyboard for iPhone roundup. This guide is about how to choose, not which one wins outright.

How six iPhone dictation options compare

The table below draws only from each product's own published pages, checked and dated. Where a page does not state a fact, the cell says "not stated" instead of guessing.

Several of these products also ship a Mac or Windows app; this table only covers what each states for iPhone.

A few patterns stand out once the six sit next to each other. Only Parlo and Spokenly name a dedicated third-party iPhone keyboard. Every product states where processing happens. Parlo answers on-device for everyone, Wispr Flow answers the cloud for everyone, and the other four vary by setting or device. Account requirements split cleanly too. Apps with a purely local free tier, such as Parlo and Spokenly's local mode, do not ask for one. Apps built around a cloud-hosted free tier tend to gate the paid plan behind sign in. Price clusters in a narrow band, roughly $8 to $15 a month, once you look past the free tiers. Their shape, more than the sticker price, usually decides whether a product fits how much you actually dictate.

Who actually needs a dedicated app

If you only dictate the occasional text message, Apple's dictation already does the job. Installing anything else adds a step for little gain. A dedicated app earns its place when at least one of three things is true. You want the raw transcript cleaned up before it lands. You want a searchable record of what you have dictated over time. Or you dictate long enough passages that filler words and false starts cost you real editing time afterward. Writers, people who answer a lot of email by voice, and anyone who prefers talking to typing tend to land in this group quickly. Someone who only ever fires off a quick reply rarely needs more than what already ships on the phone.

The keyboard asks for Full Access. What does that grant?

A keyboard that dictates, Parlo's included, asks for Full Access during setup. Full Access lets the keyboard extension write to a shared container with its companion app and reach the network. On its own it grants nothing close to microphone access. Apple keeps every keyboard extension inside a restricted sandbox, and that sandbox never includes the microphone, Full Access on or off.

That restriction is why a dedicated keyboard app cannot simply record you while you type. Parlo's keyboard hands the mic tap to its companion app instead. The app opens, records, and transcribes, then passes the finished text back through a shared container the two share. You swipe back to the field you left, and the keyboard inserts the text once. Apple requires that manual swipe; no app can skip it and insert text automatically while you are somewhere else. The full mechanics of the handoff are covered in why a voice keyboard cannot record audio on iOS and in keyboard extensions and Full Access explained.

Is a speech to text app accurate enough on iPhone?

Accuracy on iPhone comes from two layers, and it helps to know which one a product is actually selling. The first layer is the speech engine, the part that turns sound into words. The second is a cleanup pass, an optional step that edits the raw transcript. It removes filler sounds, trims false starts, and applies a spoken command such as "new paragraph." A product can be strong at one layer and weak at the other.

Products that run recognition in the cloud, such as Wispr Flow, are betting that a larger server-side model beats what a phone can run alone. Products built around Apple's on-device speech engine are betting the other way, that the newest on-device model plus a cleanup pass gets close enough without ever leaving the phone. A custom dictionary, offered by most products in the table above, teaches names and terms either way. What it actually feeds differs by product; on Parlo it is a do-not-correct list for the keyboard's autocorrect, not a change to recognition or cleanup itself. Parlo is English-first, and it is a beta, so expect rough edges there. The four named competitors each state support for 100 or more languages.

Where does your audio go?

This is worth asking about any speech to text app, since the honest answer varies more than most marketing pages suggest. Wispr Flow states plainly that transcription always happens in the cloud. superwhisper processes on device by default, and Spokenly offers local models alongside cloud ones, each with a cloud path you opt into. Monologue processes in the cloud by default with a stated zero data retention approach, and offers local transcription as a Mac-only option.

For Parlo, audio is captured, transcribed on your device, and discarded; nothing about a recording is ever stored. The first time you dictate, iOS may need to download Apple's on-device speech model, which does touch the network once. After that, transcription and cleanup both run locally on the phone. History sync is on by default once you sign in, and your transcripts, never audio, are end-to-end encrypted through your private iCloud. Watch for a phrase like "100 percent offline" on any product's page; a first-run model download, or a routine update check, is normal and does not make that phrase true.

What does an iPhone dictation app cost?

Two different pricing shapes show up across this category. Apple's dictation has no separate price; it ships with the phone at no charge. Everything else in the table above sells a free tier bounded by a word cap, a note cap, or a limited model. Next to it sits a paid plan running roughly $8 to $15 a month, depending on the product and the billing term chosen.

Parlo takes a third shape: the whole app is free during its public beta, with no word cap, no note cap, and no account required to start using it. That is a beta pricing model, not a promised permanent free tier, and it may change once the beta ends. Install Parlo for iPhone through the public TestFlight link on the beta page; Parlo for Mac installs from the same page if you dictate there too.

Where Parlo fits

Parlo transcribes with Apple's SpeechAnalyzer and SpeechTranscriber, the same on-device engine behind the newest system dictation, and it is the only engine Parlo uses. There is no cloud fallback and no second model to pick between.

An on-device language model then cleans the transcript up. Filler sounds, false starts, and repeated words come out. A length guard rejects any rewrite that comes back implausibly longer than what you actually said.

The keyboard cannot record audio, because iOS does not allow a keyboard extension to use the microphone. The app records and transcribes instead, then hands the text back to the keyboard through the shared App Group the two targets use. The keyboard itself never talks to iCloud.

Using it in practice is three steps: tap the mic key in the keyboard, speak in the Parlo app that opens, then swipe back to the field you left. The keyboard inserts the text once, only in that field, and a finished result expires after 30 seconds if you never swipe back to collect it.

Search across your past dictations is semantic and runs on-device. A keyword match comes first, then sentence-level similarity, then recency, so a half-remembered phrase can still find the right note later.

Any saved note opens for editing by hand, or with an on-device transform: turn it into an Email, a Bullets list, a shorter version, or a quick grammar fix. A saved edit keeps the original dictated words too, so a one-level revert always gets you back.

Your custom dictionary is a do-not-correct list the Parlo keyboard checks before its autocorrect changes a word you typed. It protects names and terms you have added; it does not change what the speech engine hears or what the cleanup pass does to a sentence.

The apps carry no purchase or subscription code today. Install on iPhone through the public TestFlight link, and the same beta page carries the Parlo for Mac install, with no account needed for either one.