CloudEffects

Add-on — Voice capture

You explain things best when you are talking.

Seminars, interviews, meetings, conversations with your veterans. What you say is more concrete than what you write — and it sits asleep in recordings. Hand over the recording you already made, and it is turned into text and — as a rule, once you adopt it — into your company’s knowledge.

Works · $19/mo + credits

When you need it

When it helps.

Anywhere people explain your business out loud.

Keeping a veteran’s know-how

What you heard from them before they retire becomes the company’s knowledge, not one person’s memory.

Turning a seminar into an asset

What you said on stage becomes material for the next article and for the answers the AI gives.

Keeping what was said in a meeting

Hand over the recording. If the meeting tool already made a transcript, paste that instead.

Hearing your customers’ own words

Interviews and consultations hold the words customers really use — the ones your website should use too.

Devices and files

Record with what you already use.

OS iXanagi takes the recording after it is made — you do not need a new device or app. There are four ways in: upload an audio file, give the address of an audio file, paste a transcript, or upload a transcript file.

Your phone’s voice recorder

The recording app on the phone in your pocket. Upload the file it saves (m4a and others).

A voice recorder

Copy the recording to your computer and upload it — mp3, wav and the other common formats.

An online meeting’s recording

A recording saved by a meeting tool (mp4, webm and others). If the tool exported a transcript (VTT), that works too — with no transcription cost.

An AI voice recorder’s cloud

Notta, Otter and a generic format can send recordings in by webhook once the brand allows outside recorder clouds; PLAUD only when approved by name. By default, only files handed over in the console are accepted.

A transcript you already have

Paste plain text, or upload a .txt / .srt / .vtt file. Timestamps are taken out of the text.

An audio file on the web

Give the address of an audio file, and it is fetched and transcribed. Addresses inside private networks are refused.

How it flows

From a recording to your knowledge.

As a rule, knowledge is added once you adopt it. The exceptions are items whose basis is confirmed within the material itself, and imports where you chose to adopt everything at once — in both cases the reason stays on record. Each step below stops where it should.

  1. Record, then hand it over

    Record with your usual device. Upload the file in the console and say whether everyone agreed to the recording, and where it was made.

  2. Consent checked, then transcribed

    Where everyone’s consent is treated as required, transcription waits until it is confirmed. An audio file you hand over is deleted right after it becomes text.

  3. Held back, then sorted

    A transcript containing fixed-format personal data (email addresses, phone numbers, card numbers and the like) or medical words such as diagnosis or prescription is encrypted and waits for a person. Each one is compared with your concepts and lands in an inbox, relevant ones marked.

  4. Approved, it becomes knowledge

    You choose what to take in. Knowledge candidates are drawn from it, and, as a rule, the ones you adopt join your company’s knowledge.

What changes

Until now, and from now on.

What changes when what was said is kept as knowledge.

What protects you

Built to record safely.

A recording holds other people’s voices. These rules work whether or not anyone remembers them.

No consent, no transcription

If you say the recording was made in Japan, the EU or one of 13 US states such as California — or do not say where — nothing is transcribed until everyone’s consent is confirmed. Elsewhere, a record of consent to the recording is still required.

The audio is not kept

An audio file you hand over is deleted right after it is turned into text; only the text remains. When the audio comes from a web address or an outside recorder cloud, the original stays there.

Personal data waits for a person

A transcript containing fixed-format personal data (email addresses, phone numbers, card numbers and the like) or medical words is encrypted, hidden in lists, and goes no further until someone checks it. Names and addresses, which have no fixed format, may not be caught.

Nothing becomes knowledge on its own

Transcripts wait in an inbox. As a rule, only what a person takes in and adopts joins the company’s knowledge. The exceptions are items whose basis is confirmed within the material, and imports where you chose to adopt everything at once — the reason stays on record either way.

Where the audio lives is your policy

Each brand chooses whether recordings from outside recorder clouds are accepted at all. The default accepts only what you hand over yourself.

Can be erased on request

A capture can be deleted completely, for example when someone asks for their data to be removed.

Price

What it costs.

A small monthly fee, and credits only for the minutes that are transcribed.

$19 / month

The add-on’s monthly fee.

1 credit per minute of audio

Counted from the actual length the transcription service reports, per started minute — not guessed from the file size.

Transcripts cost 0 credits

Pasting or uploading a transcript uses no transcription service, so it uses no credits.

Your own OpenAI key: 0 credits

Transcription uses OpenAI’s Whisper. If you register your own OpenAI key, it is billed to you by OpenAI and uses no credits.

What it needs

What it takes

Every App comes with the plan. Unless an App is activated, its capabilities cannot be used — from the screen or the API.

Apps it needs

No extra App is needed.

It runs on Knowledge, which comes with the plan.

Also needed

Recording consent from everyone taking part

Where everyone’s consent is treated as required, transcription waits for it.

Also needed

A recording or a transcript

From the device you already use.

Where this stands

Files Audio: mp3, m4a, wav, webm, mp4, mpeg, mpga, ogg and flac, up to 10 MB per file from the console. Transcripts: plain text, .txt, .srt and .vtt, up to 2 MB.

How it records OS iXanagi takes recordings after they are made; it does not record in the browser, transcribe live, or tell speakers apart. Transcription goes through OpenAI’s Whisper; OS iXanagi does not run its own speech server. Not used for training No recording or transcript is used to train an AI model, by us or by the transcription provider. Only what you adopt becomes knowledge.

Works well with

Better together.

App

Knowledge

Where captured voices become the company’s knowledge.

Learn more

Connector

YouTube

Captions from your own channel become knowledge the same way.

Learn more

Connector

Website chat

Answers customers from the knowledge you captured.

Learn more

Connector

LINE

The same knowledge answers on LINE too.

Learn more

How would it work for your business?

Curious? Ask the AI directly — it answers in terms of your business.

You explain things best when you are talking. | CloudEffects