Live and file AI translation
Four steps, all on your own systems.
Tolkr™ listens, understands, translates and speaks, live or from a file, all on your own systems.
Recorded from Tolkr™ Studio. A press statement, 
French to 
English.
- 01Sound to voiceone sentence at a time
- 02Live or a filethe same steps, both ways of working
- 03On your systemsevery step, nothing sent out
The four steps
From sound to a voice.
The same four steps, live or from a file. Switch to see what changes.
-
01
Listen
The whole PC, one app, a microphone or a file. A quiet speaker is still heard, and Tolkr™ follows the speech sentence by sentence.
The whole file, from start to end, followed sentence by sentence.
-
02
Understand
Each sentence is written in its own language, and what is still being heard shows grey on screen. The language can be found by itself.
What is said is written down, and the speakers are told apart.
-
03
Translate
Sentence by sentence, in the context of the conversation, so a reply still makes sense. Unsure lines are marked.
Sentence by sentence, in the context of the conversation, and right for who speaks to whom.
-
04
Speak
A natural voice speaks the translation as the speaker goes on, and every line stays in the captions.
Each speaker gets a voice of their own, every sentence starts with its original, and the programme comes back mixed.
Checks
What keeps the text right.
Tolkr™ shows and speaks what was said, and tells you where to look twice.
- Only speechNever noise
Background noise does not turn into words.
- Every lineKept
Every line is saved with its sound, to check against the original.
- UnsureMarked
A line Tolkr™ is unsure of is marked, so you know which one to check.
From a file
Every sentence fits its time.
A translation is often longer than its original. Tolkr™ fits each one into the time of its original, and the sentence list shows every one.
About Dub a file- StartsWith its original
- EndsBefore the next sentence begins
- MixedWith the original lowered under the voice
- ReportedEvery sentence, in the sentence list
Where it runs
On your systems, sized to it.
Tolkr™ sets itself up for the workstation.
- Graphics cardWhen there is one
With one, Tolkr™ is at its fastest.
- ProcessorAlways possible
Without one, Tolkr™ runs on the processor.
- Your choiceAny time
You can change the setup at any time.
Asked first.
Not answered here? Ask the people who build it
Does it translate word by word?
No. Tolkr™ translates whole sentences, so the meaning comes out right.
Does it work with two or more speakers?
Yes. In a dub, each speaker gets a voice of their own.
Does it need the internet?
No. Every step runs on your own systems.
Can it find the language by itself?
Yes, live it can find the language of each sentence.
See the steps on your own sound.
We run Tolkr™ on a recording or a stream you choose.
Request a demo