Sawarni combines an iPhone teleprompter, camera and video editor with a specific captioning model: the words come from the user's script and speech sets their timing. That is different from independently transcribing everything actually spoken. Its official description emphasizes Arabic and Darija alongside French and English, with paragraph-level text direction for mixed-language scripts.
This First look is based on the official listing. We have not recorded a take, checked voice following or exported a video. Recording, editing and the internal library are described as free, while each save to Photos and each share requires a short advertisement.
Based on official store information. We have not tested this app hands-on.
What it does
The script sits near the front camera, with adjustable font size, height and background. Follow My Voice moves the text during speech and waits during pauses, while a words-per-minute option provides steady scrolling. A mirror mode is listed for physical teleprompter rigs. We have not tested recognition, compatible hardware or how a skipped line affects position.
The camera records vertical video at 1080p with sound. Front and back cameras and the framings available on the actual iPhone are supported, rather than every phone being promised identical zoom options. Mirroring can be enabled or disabled, and a three-step countdown precedes recording. Check the actual framing and recorded result before treating the local preview as a complete rehearsal.
Editing includes cutting mistakes, trimming, splitting, adding text and undo. Captions have three styles and editable lines. The source says script-derived wording preserves chosen spelling, including Darija. That can be useful for an intentional phrase, but it also means a script containing an error or a line not delivered needs review rather than being treated as an accurate transcript by default.
Takes remain in the in-app library, with edits associated with each video and no account required. The description does not establish a complete backup, cross-device synchronization or recovery policy. Keep important original text and media independently if their preservation matters beyond the current device.
Before you install
The main editorial check is comparing the spoken take with the script-derived captions. A presenter can improvise, omit a phrase or correct a number while speaking. Preserving the script's spelling is not proof that the caption now matches what the audience hears. We have not measured alignment or the behavior of an unscripted passage.
Mixed text direction also needs actual inspection. A paragraph containing Arabic and another in French can have different reading needs, while names, punctuation and line breaks may still need correction. The listed languages should not be expanded into a promise of all-language transcription or automatic translation.
Voice following and caption timing are separate uses of speech. Neither is evidence that the app fact-checks the content, diagnoses pronunciation or guarantees natural eye contact. Rehearse a representative passage and inspect the received video rather than assuming one successful sentence validates the whole recording.
The ad condition matters to a client deadline. A take can exist in the library while saving or sharing still requires an advertisement. Confirm the actual flow before depending on an immediate handoff. The reviewed description does not establish that watching one advertisement permanently unlocks all future exports or that an ad is always available under every network condition.
We have not audited camera, microphone, advertising or storage behavior. A no-account statement is not the same as a complete local-only privacy guarantee. Check current permissions and terms before recording confidential material or another person, and understand the separate handling of a deliberate share to another service.
For a client introduction, keep names and promised deliverables in the approved script before recording. An off-script clarification may be audible without becoming new caption text. That distinction should guide the review: compare the actual spoken message with the prepared words rather than treating correct spelling as evidence that everything said is represented.
Price and access
The linked US App Store offer is a free download. The description includes recording, editing and the internal library without payment, with a short ad unlocking each individual save to Photos and each individual share.
No subscription or numeric upgrade price is described in the reviewed text. Do not invent an ad-free paid tier or portray export as unconditionally free of another step. We have not tested the ad gate, repeated exports, billing screens or access after a network interruption. The internal editing entitlement and an external deliverable are different workflow states.
Our take
Sawarni's strongest proposition is a script-centered recording workflow with language-direction support and intentionally preserved caption wording. That can be relevant to a creator who prepares the exact message in advance and is willing to review whether the delivered take still matches it. The per-export ad gate is part of that practical decision.
Begin with a short mixed-direction or otherwise representative script, record it and compare the words, timing and final output. Test the actual saving and sharing path before using it for a deadline. Our take remains source-based: language behavior, voice tracking, caption alignment and privacy have not been independently tested, and script captions should never be mislabeled as a verified transcription.
Sources
App icon from the official store listing.
