Skip to content
DownloadRU

Products

All products and statuses →EcosystemHelpDownload

RecUmi Voice

In development · technical stage

Say it, and it will not be forgotten: a note, a meeting or a lecture becomes text.

RecUmi Voice is a standalone product of the RecUmi ecosystem: a personal voice recorder that saves a recording with marks on important moments and turns it into text right on the phone. The product is at a technical stage of development: it cannot be downloaded, and there is no timeline or release date.

Everything described below as "Implemented, in testing" works in the test version and is still being checked. The test version is not published.

RecordingMarksTextSummary one tap"Record" "Markmoment" on the phone,no cloud key points, decisions,tasks: planned
The recording path: three steps are in testing, the summary is a planned capability.

Who it is for and why

Voice is faster than a keyboard, but a recording that cannot be quickly found and read usually stays unheard. Voice makes it possible to return to what was said: by a mark, by the text, by the title.

Managers

The outcome of a meeting and what was agreed, right after the conversation, while nothing is forgotten.

Specialists on site

Observations on site by voice, when your hands are busy and typing is awkward.

Students and listeners

A lecture, lesson or webinar with marks on important places and text, so you can revise.

Every day

A thought, an idea or a to-do list: say it on the go, read it later.

What RecUmi Voice can do
Implemented, in testing the feature is implemented in a test or internal version and has not yet been made available to users as a finished public product Planned designed, development ahead Idea, discussed earlier but not yet approved for development not confirmed, not promised

One-tap recording

Implemented, in testing

After the first setup, recording starts as soon as you tap "Record".

How it is used
Recording continues if you minimise the app or turn off the screen; while it runs, Android shows a notification and the microphone indicator.
What data
Audio recording in the app's internal storage.
Benefit
No need to look for a recorder and set it up before every note.

Pause on a call

Implemented, in testing

If you answer a phone call or a messenger call during recording, the recording is paused.

How it is used
Recording does not resume by itself after the call.
Benefit
The call does not end up in your note.

Mark a moment

Implemented, in testing

The "Mark moment" button puts a mark at the current place in the recording; all marks are shown as a list.

How it is used
During a lecture or a meeting, mark what is important, then jump to it with one tap.
What data
The time of the mark within the recording.
Benefit
No need to relisten to an hour of recording for a single phrase.

Recording library

Implemented, in testing

A list of recordings and a card for each one with a player. You give each recording its name.

What data
Title, start time, duration, marks.

Transcription on the phone

Implemented, in testing

A recording is turned into text on your command. Recognition runs on the phone; cloud recognisers are disabled in RecUmi Voice.

How it is used
It uses the built-in Vosk engine with a model for the language you need, or the Android system recogniser if the system confirms that it works on the device.
What data
The text is stored on the phone together with the recording.
Benefit
You can read and copy the text instead of relistening.

Recognition languages

Implemented, in testing

Language models are downloaded only when you tap: before downloading, the source and size are shown, and afterwards the app checks the checksum.

How it is used
You can choose to download over Wi‑Fi only.
Benefit
With a downloaded model it works without the internet; for languages recognised by the Android system this depends on the device and its language components.

Copying text

Implemented, in testing

The transcript text can be copied and pasted into an email, a messenger or a document.

ZIP backup

Implemented, in testing

Recordings can be saved to a ZIP file in a place you choose.

How it is used
The backup contains the audio and information about the recordings (title, time, duration, marks); transcript texts are not included. The file is not encrypted.
Benefit
Recordings can be moved or kept separately from the phone.

Editing the transcript

Implemented, in testing

Correcting mistakes in the transcript text; the original text is kept and you can return to it.

How it is used
Known limitation of the test version: returning to the original discards all edits.
Benefit
Text you can trust, without losing the original.

Search and jump to a moment

Planned

Search across the text of all recordings and jump from a word to the place in the audio.

Benefit
Find the right place by a word, not by memory.

Summary and key points

Planned

The main points of a long recording in a few lines: key points, decisions made, agreements.

How it is used
This is a future feature; we will announce it only when it is ready and tested.
Benefit
The outcome of a meeting without listening again.

Tasks and reminders

Planned

Voice → text → task suggestion → your confirmation → done.

How it is used
A task or reminder appears only after your confirmation.
Benefit
What was agreed does not stay only in the recording.

Study scenarios

Planned

A lecture with section marks, text and a short summary for revision.

Benefit
Exam preparation from your own recordings.

Generative AI for summaries

Idea, discussed earlier but not yet approved for development

Full understanding of a recording by a neural network.

How it is used
We do not promise this now: the technology and its accuracy on a phone are not proven.

How it works

From the tap to the text in four steps. Everything happens on your phone.

  1. 1. RecordOne tap: recording runs in the background too
  2. 2. Mark"Mark moment" at an important place
  3. 3. SaveThe recording with its title and marks is in the library
  4. 4. ReadText on the phone on your command

The fifth step, summary and tasks, is a planned capability.

Scenarios
TodayThe meeting is over, and what was agreed is in the memory of three people, each with their own version.
With RecUmiA meeting recording with a mark at every decision; afterwards, text from which the agreements are visible in a minute.
Meeting
TodayThe lecture is recorded in full, and finding the right topic means listening for an hour straight.
With RecUmiMarks at every new topic of the lecture and text: you can jump straight to the right place.
Study
TodayOn site, an important observation was put into the phone's notes in a single line.
With RecUmiThe observation is dictated in detail with marks; the text can be copied into a report.
Site visit

Part of the ecosystem

Voice is for personal voice recordings. Phone calls are recorded by RecUmi Calls. Together with the other products, Voice will pass agreements on to where they are worked with.

  1. VoiceA recording of a meeting with a customer
  2. VoiceText and key decisions
  3. CRMThe next action in the customer record

Planned The link between Voice and CRM will appear together with the release of the products and the shared RecUmi Core foundation.

All features and their stage

TASK 01Section

Recording

Features: 5 · implemented, in testing: 3 · planned: 2
  • One-tap recording, in the background, with a notificationImplemented, in testing
  • Pause on a call; recording does not resume by itself after the callImplemented, in testing
  • "Mark moment" and a list of marksImplemented, in testing
  • "Stop and save" right from the notificationPlanned
  • Long recordings and work by fragmentsPlanned
TASK 02Section

Library

Features: 2 · implemented, in testing: 1 · planned: 1
  • List of recordings and a card with a playerImplemented, in testing
  • Automatic recording titlesPlanned
TASK 03Section

Text

Features: 6 · implemented, in testing: 4 · planned: 2
  • Transcription on the phone (Vosk or the system recogniser on the device)Implemented, in testing
  • Downloading language models on tap, Wi‑Fi only as an optionImplemented, in testing
  • Copying textImplemented, in testing
  • Transcript editing with the original preservedImplemented, in testing
  • Text search and jumping from a word to the moment in the recordingPlanned
  • Timecodes and text fragments, punctuationPlanned
TASK 04Section

Storage

Features: 3 · implemented, in testing: 1 · planned: 2
  • ZIP backup to a place you chooseImplemented, in testing
  • Export and sharing of the recording and its textPlanned
  • Cloud copy, if you choosePlanned
TASK 05Section

Interface

Features: 1 · implemented, in testing: 1
  • Interface language and theme in settingsImplemented, in testing
TASK 06Section

Summaries

Features: 4 · planned: 3 · Idea, discussed earlier but not yet approved for development: 1
  • Summary, key points, decisionsPlanned
  • Tasks and reminders with your confirmationPlanned
  • Generative AI for summariesIdea, discussed earlier but not yet approved for development
  • Topics and key points of a recordingPlanned
TASK 07Section

Study

Features: 2 · planned: 2
  • A lecture summary with sectionsPlanned
  • Definitions, review questions and self-check cardsPlanned
TASK 08Section

Ecosystem

Features: 1 · planned: 1
  • Passing agreements on to RecUmi CRMPlanned
TASK 09Section

Languages

Features: 1 · planned: 1
  • Translation of the recording textPlanned
TASK 10Section

Organisation

Features: 1 · planned: 1
  • Task deadlines and calendarPlanned

"Implemented, in testing" means the feature is implemented in a test or internal version and has not yet been made available to users as a finished public product. Where: in the RecUmi Voice test version; the test version has not been published. This is not a live RecUmi service.

Where recordings are stored

On your phone

  • Recordings, texts and marks are in the app's internal storage.
  • The app does not send them to the developer.
  • No registration, ads, analytics or crash-reporting services.

What goes to the internet

Only the download of language models, when you tap, over a secure connection from the Vosk project website. Recordings and texts are not sent.

The app does not need the call log, contacts, SMS or location.

Details: the RecUmi Voice privacy policy.

Development stages

Done

  • Test version: recording, pause on a call, marks, library, transcription on the phone, ZIP backup

Next

  • Transcript editing, text search, jumping from a word to the moment in the recording

Later

  • A short summary, key points and tasks with your confirmation
  • Lecture summaries
  • Connection with RecUmi CRM

There are no dates: the stage changes when a feature has been verified.

How to connect

You set up yourself

  • Download the recognition language models
  • Choose to download over Wi‑Fi only
  • Save a ZIP backup to a place you choose

Through an external provider or Android

  • Android permissions: microphone and notifications
  • The Android system recogniser, if it works on the device

Not available yet

  • App download: the product is at a technical stage
Technical details

Recognition uses the built-in Vosk engine with a model of the chosen language, or the Android system recogniser on the device; cloud recognisers are turned off. Models are downloaded from the Vosk project site over HTTPS after you tap; the app shows the source and size and verifies the checksum. Details are in the RecUmi Voice privacy policy.

Frequently asked questions
Can I download RecUmi Voice?

No. The product is at a technical stage of development; there is no download button. We will announce a release only if it happens.

When will it be released?

There are no dates. We do not name dates until a release has happened.

How is Voice different from Calls?

Voice is for personal voice notes, meetings and lectures. Phone calls and WhatsApp calls are recorded by RecUmi Calls; Calls also has a simple personal voice recorder.

Does Voice record phone calls?

No. During a call, recording is paused and does not continue on its own.

Do I need the internet for transcription?

Only to download a language model once. With a downloaded model, the text is created on the phone without the internet; for languages recognised by the Android system this depends on the device and its language components.

Can I fix the text?

In the test version, yes: the transcript text can be edited and the original is kept.

How much space do recordings take?

Recordings and language models are stored in the phone's memory; the model size is shown before you download it.

How accurate is the transcription?

Accuracy depends on recording quality, language and device. Check important passages against the audio.

Will there be AI summaries and tasks?

These are planned capabilities. We will announce them only when they are ready and tested, and a task will appear only after your confirmation.

Voice is at a technical stage

RecUmi Calls is available now: call recording and a personal voice recorder for Android. Voice will appear on this page when a release happens.

RecUmi Calls: availableAll productsContact us

Related products