LipConfirm
Fix typing mistakes without speaking up.
A quiet whisper—or a silent lip cue—gives autocorrect another clue.
LipConfirm combines what you type with whisper-level audio or mouth-movement cues to help resolve garbled or ambiguous words. Typing stays primary in this correction workflow; the extra cue does not have to identify the intended word on its own.
All 30 claims allowed · U.S. Patent No. 12,737,540 scheduled to issue September 15, 2026.
Try our Interactive Demo
Intended text
restaurant
Mistyped text
Correction
How the correction is chosen
Two distinct correction strategies
Generate candidates from typing first, then use lip or whisper evidence to rescore and select.
- 1. Candidates
- 2. Evidence
- 3. Scores
- 4. Commit
Candidate ranking
Illustrative scores · 0–100Typing only · no auxiliary evidence
Type the word to see its candidates.
Waiting for the mistyped word.
Illustrative simulation. No microphone or camera is accessed.
Typing stays primary in this simulation. Whisper and lip cues guide correction rather than provide a standalone dictation or lipreading result. A single articulation can supply both cues in a combined implementation; combined capture is not simulated here.
Why another cue helps
Even LLM-informed autocorrect can struggle with garbled or ambiguous typing. LipConfirm adds evidence from you—a quiet whisper or silent lip cue—to help identify the word you intended.
A cue can come before or after the mistake.
Quietly articulate a difficult word as you type it, or if a word comes out wrong, add a cue afterward and LipConfirm reconsiders it using the typing you already did.
Whisper-assisted correction
Whisper assistance needs no camera. The letters you type help narrow the possible corrections, so the whisper does not have to identify the word on its own.
Lip assistance, on request
Optional lip assistance can use a brief, requested capture rather than watching throughout typing. A single quiet articulation can also provide both audio and visible mouth movement.
How it works
Preliminary desktop demos are available for both lip-assisted and whisper-assisted correction. The videos below show lip-assisted correction.
Capture useful cues
With permission, the microphone can remain available during typing for occasional whisper-level cues. Optional camera capture can be limited to requested lip assistance. The architecture supports local processing and discarding raw audio and frames after use.
Use evidence inside correction
Whisper-level audio or lip evidence is evaluated together with the typed input, rather than having to resolve the word independently first. It can rescore or validate candidates, help recover missing candidates, or enter a learned multimodal model together with text.
Suggest, confirm, or replace
The combined evidence can update a suggestion, confirm a word, or support automatic replacement. When confidence is insufficient, the system can leave the text unchanged or present alternatives rather than forcing a correction.
What the demos show
- Messy input, clear intent. Selected examples in which the text-only baseline misses or misranks the intended correction.
- Typing stays primary. Lip cues in the videos guide a typed correction rather than independently transcribe the mouthed word. Preliminary desktop whisper demos apply the same principle using quiet audio.
- No need to repeat everything. An optional cue can accompany a difficult word. The user does not need to whisper or mouth every word they type.
- Note: The camera panel is a behind-the-scenes visual aid. An integrated keyboard need not display the feed, and requested visual correction need not keep the camera running throughout typing.
Preliminary desktop demos are available for both lip-assisted and whisper-assisted correction. The videos below show lip-assisted correction.
What the demos show
- Messy input, clear intent. Selected examples in which the text-only baseline misses or misranks the intended correction.
- Typing stays primary. Lip cues in the videos guide a typed correction rather than independently transcribe the mouthed word. Preliminary desktop whisper demos apply the same principle using quiet audio.
- No need to repeat everything. An optional cue can accompany a difficult word. The user does not need to whisper or mouth every word they type.
- Note: The camera panel is a behind-the-scenes visual aid. An integrated keyboard need not display the feed, and requested visual correction need not keep the camera running throughout typing.
Built around ordinary typing
Keep typing normally and add an occasional quiet cue for a difficult word. LipConfirm is designed to bring that evidence into existing keyboard, autocorrect, spell-checking, and language-model infrastructure—not replace the typing workflow.
Technical demonstrations and IP discussions available. Preliminary desktop demos cover both whisper-assisted and lip-assisted correction.
Patent & IP
All 30 claims allowed. U.S. Patent No. 12,737,540 is scheduled to issue September 15, 2026. The allowed claim set addresses lip-assisted correction, shared multimodal operation, and a camera-free whisper-assisted pathway that generates textual proxies for an existing spell-correction engine.
Two U.S. continuation applications are pending, seeking additional protection for lip- and whisper-assisted autocorrect. One has been granted Track One prioritized examination.
PCT/US2026/035963 is pending. The USPTO, acting as the International Searching Authority, issued a written opinion with positive novelty, inventive-step, and industrial-applicability findings for all 30 claims. The international search report contains no category X or Y citations.
FAQ
Why not just use dictation?
Dictation is useful when it recognizes the speech and fits the setting. LipConfirm focuses on difficult typed words: it evaluates whisper-level evidence against text-supported candidates, so the audio need not identify the intended word independently. The goal is to make quiet, uncertain cues useful for correction—not replace dictation when it works well.
What is whisper-level audio?
Whisper-level audio is quiet acoustic evidence, including low-amplitude or breathy speech. LipConfirm evaluates those clues together with typed input; a complete, correct transcript is not required. Usable volume depends on the microphone, distance, background noise, and model.
Does LipConfirm require a camera?
No. Whisper-assisted correction combines microphone input with typing and does not need a camera. Optional lip assistance can provide a silent cue or reinforce a quiet utterance.
Would the camera run throughout typing?
Continuous camera use is not required. In the on-demand approach, the camera stays off until lip assistance is requested, captures a short clip of the word, then stops. Longer visual sessions remain an optional, user-enabled configuration.
Can the microphone stay available while I type?
Yes, in an opted-in typing session. The microphone can remain available so an occasional quiet cue accompanies a difficult word as you type it. Alternatively, it can activate only when assistance is requested. There is no requirement to whisper every word.
Can I correct a word after it appears?
The architecture supports associating a new cue with the word just typed and reconsidering the correction using the retained typing evidence. Optional lip assistance can use a brief capture requested for that word.
Can whisper and lip cues work together?
Yes. One quiet articulation can supply audio and lip cues. The architecture combines both with typed evidence and adjusts their influence according to reliability. An unavailable or unreliable channel can contribute less or be omitted.
What is demonstrated today?
Preliminary desktop demonstrations are available for both whisper-assisted and lip-assisted correction. The interactive website demo is an illustrative simulation with preset examples and scores; it does not access your microphone or camera. It explains the correction strategies rather than measuring live recognition accuracy.
Does this require the internet or cloud processing?
LipConfirm is designed to support on-device correction without sending raw mouth-region frames or audio to the cloud. Raw sensor data can be discarded after feature extraction or scoring. Deployment choices depend on the host platform and model.
Will this drain battery or slow down typing?
The design supports brief camera use and session-based or requested microphone capture. Actual battery use and latency depend on sensor settings, recognition models, and the device.
How does the extra evidence enter autocorrect?
It can rescore or validate existing candidates, help generate or recover candidates, or enter a learned model together with text. A proxy-string pathway can turn speech-related clues into spelling-like inputs for an existing correction engine. The interactive demo illustrates late and early fusion as two distinct strategies.
Do I need to perfectly mouth the words?
Text-assisted correction can use partial cues, such as syllable patterns and mouth shapes, rather than independently recognizing every word from lips alone. Useful visual evidence still depends on articulation and capture quality. Optional personalization can adapt scoring to a user.
How much accuracy improvement does LipConfirm provide?
The desktop demonstrations show selected correction scenarios, not a population-wide accuracy figure. Real-world benefit depends on the implementation, capture conditions, and user behavior. Reliable operation at very low volume and normal phone-typing distance remains a mobile validation target.
Is LipConfirm only for mobile phones?
The architecture supports keyboard-based correction on phones, tablets, desktop systems, wearables, and extended-reality devices. Each platform can use the available, user-permitted combination of typed, acoustic, and visual evidence.
Will I still see suggestions?
A high-confidence result can be inserted or confirmed automatically. When the evidence is uncertain, the system can offer alternatives, retain the current text, or wait for another cue. The user does not need to see the internal candidate scores or camera feed.
Is this patented?
All 30 claims of the U.S. parent application have been allowed, and U.S. Patent No. 12,737,540 is scheduled to issue September 15, 2026. Two U.S. continuation applications and a PCT international application are pending. One continuation has been granted Track One prioritized examination. See Patent & IP for details.
Contact
Email: hello@lipconfirm.com
LipConfirm was conceived and developed by Andre Persidsky, a multidisciplinary engineer and inventor with experience building, patenting, and licensing technology. Andre has previously licensed technology to commercial partners including Spotify, and has led hardware/software health-technology projects in collaboration with external research institutions.