Realtime translation

Speak one language,write in another.

Two synchronous language lanes. Recognized text on the left, translation on the right. Translation runs through managed cloud models — availability is per language pair.

Available where enabled

AIspeaky is currently in beta and free to use within beta limits. We do not read or store the content of what you dictate.

Two lanes · EN → ESLive
Source (EN)
"Thanks for the update — I'll send the revised proposal after lunch."
Target (ES)
"Gracias por la actualización — te envío la propuesta revisada después de comer."
Insert mode
Target language into the focused field.
Source stays visible in the tray for reference.
Language pairs currently in beta: EN ↔ ES, DE, FR, RU. Others expanding.
How it works

Recognized once, translated in parallel.

AIspeaky doesn't re-record — the recognized transcript is the input to translation.

  1. 1

    Speak in your source language

    Push-to-talk starts a session. Managed recognition returns the transcript in the source language.

  2. 2

    Translate through managed models

    The transcript is sent to a translation model for the pair you selected. Result streams back within seconds.

  3. 3

    Insert the target language

    The translated text lands in the active field, or you can copy both lanes. Source stays available in the tray.

In the product

Not every pair is equal — we say which is which.

Beta translation is a real feature, but quality varies by pair and by direction. Tags reflect current beta observations, not aspirations.

ENES
Beta quality solid
ENDE
Beta quality solid
ENFR
Beta quality solid
ENRU
Beta — style varies
RUEN
Beta — style varies
OtherOther
Not enabled yet
What's in and what's not

What's on offer, and what's not.

Translation is a cloud feature. Enable it only when the workflow benefits from it.

Included in beta
  • Push-to-talk translation with two visible lanes
  • Insert-target-language into the focused field where supported
  • Copy source + target as a single block
  • Per-pair beta quality tags shown before you send
Not yet — expanding
  • Offline translation
  • Simultaneous multi-target output (one target per session)
  • Voiceover / spoken translation playback (separate roadmap item)
  • Guaranteed idiomatic quality across every pair
Questions before install

Answered plainly.

Does audio leave my device?

For managed recognition, yes — the audio goes to the STT provider, and the resulting transcript goes to the translation provider. See the data-flow table.

Can I use it offline?

No. Translation quality in the beta depends on managed cloud models. Offline is not on the near-term list.

Do beta usage limits apply to translation?

Yes. Translation counts against the same beta usage envelope; the current limits are shown on the download and pricing pages.

Get the beta

Free during beta. Limited usage, honest expectations.

Choose your platform on the download page. Beta terms and known limits are shown before the file downloads.