Our model for the way clinicians speak.

Read the paperSee the ranking

1st of 12for finished text and medical terms

Sentences from the 44ai-clinical-de test set, exactly as Klara wrote them.

Off-the-shelf speech models aren't built for clinical work.

In the GP practice, on the ward and in the emergency department, people speak medicine, often in dialect. And when they dictate, they say the punctuation and line breaks out loud. General-purpose speech recognition is built for everyday language.

  1. GP practice

    Consultation notes, dictated between two patients

  2. Ward

    The progress note, dictated on the ward round

  3. Emergency department

    The initial assessment, once the patient is stable

That's why we built Klara ourselves.

One model for clinical speech.

Hear for yourself.

Press play. Both systems were given the same clip from the test set. Here is what each one wrote, with the red pen marking what a clinician would still have to correct before signing off.

0:00

Kidney replacement therapy, dictated in Standard German

Compare with
Our model

Klara 44ai

misheardheard correctly, but not formatted as in the report

Four examples from the test set. Full results in the paper and on GitHub.

Less red ink.

We ran twelve speech recognition systems on three clinical dictation test sets and compared every transcript with the finished report.

The next-best system still leaves 20 corrections for every 100 words and punctuation marks. Klara leaves fewer than half as many.

  1. 9per 100

    Klara

    44ai

  2. 20per 100

    Corti Symphony

    Corti

  3. 23per 100

    Scribe v2 Medical

    ElevenLabs

  4. 25per 100

    gpt-4o-transcribe

    OpenAI

  5. 28per 100

    Dragon Medical One

    Microsoft

  6. 28per 100

    Scribe v2

    ElevenLabs

  7. 30per 100

    Nemotron 3.5 Transcribe

    NVIDIA

  8. 34per 100

    MAI-Transcribe-2

    Microsoft

  9. 44per 100

    Cohere Transcribe

    Cohere

  10. 46per 100

    stt-async-v5

    Soniox

  11. 47per 100

    Whisper large-v3

    OpenAI

  12. 57per 100

    Nova-3

    Deepgram

Formatted accuracy, %. The system's output compared with the finished report, including punctuation and formatting.

  1. 1Klara44ai90.9
  2. 2Corti SymphonyCorti80.1
  3. 3Scribe v2 MedicalElevenLabs76.7
  4. 4gpt-4o-transcribeOpenAI74.9
  5. 5Dragon Medical OneMicrosoft72.2
  6. 6Scribe v2ElevenLabs72.1
  7. 7Nemotron 3.5 TranscribeNVIDIA69.6
  8. 8MAI-Transcribe-2Microsoft66.0
  9. 9Cohere TranscribeCohere55.8
  10. 10stt-async-v5Soniox53.8
  11. 11Whisper large-v3OpenAI53.3
  12. 12Nova-3Deepgram42.6

Full evaluation in the paper and on GitHub.

Everyone is good at textbook German.

On the ward, nobody speaks textbook German. Speech is mumbled, in dialect and unstructured.

5060708090100Klara 90.7gpt-4o-transcribe 81.7Scribe v2 Medical 81.3MAI-Transcribe-2 80.3Scribe v2 80.0Corti Symphony 77.0Soniox 74.9Cohere 70.5Nemotron 69.8Whisper 69.3Dragon Medical One 54.1Deepgram 51.7Standard GermanNon-native speakersNon-nativeGerman and Austrian dialectsDialectsSwiss dialectsDialects
Word accuracy in % on 44ai-clinical-de, by speaker group.

Check it for yourself.

The scoring code, the normalisation rules and every transcript from all twelve systems are public.

$ git clone https://github.com/44ai-labs/44ai-klara-evaluation
$ cd 44ai-klara-evaluation
$ make install
$ make score

## All three test sets, weighted by audio duration

| System                       | Formatted | Word | Medical |
|                              |           |      |   terms |
| ---------------------------- | --------- | ---- | ------- |
| Klara                        |      90.9 | 93.9 |    90.4 |
| Corti Symphony               |      80.1 | 89.9 |    89.6 |
| ElevenLabs Scribe v2 Medical |      76.7 | 90.7 |    89.0 |
| gpt-4o-transcribe            |      74.9 | 90.9 |    88.5 |
| Dragon Medical One\*         |      72.2 | 80.4 |    76.7 |
| ElevenLabs Scribe v2         |      72.1 | 90.0 |    87.8 |
| Nemotron 3.5 Transcribe      |      69.6 | 87.3 |    81.5 |
| MAI-Transcribe-2             |      66.0 | 90.6 |    84.6 |
| Cohere Transcribe\*†         |      55.8 | 82.5 |    78.1 |
| Soniox stt-async-v5†         |      53.8 | 88.0 |    83.6 |
| Whisper large-v3\*†          |      53.3 | 82.1 |    78.3 |
| Deepgram Nova-3†             |      42.6 | 80.0 |    73.3 |
PaperKlara: Medical Speech Recognition for Real-World Clinical EnvironmentsRead the paper

Top quality, full data sovereignty.

You decide where Klara runs: on your own servers or in a sovereign cloud. Both with the same quality.

  • On premises, in your own data centre
  • Sovereign cloud
Visit the Trust Center
  • Medical device
  • ISO 13485:2016
  • EU AI Act
  • EU GDPR
  • No US CLOUD Act
  • ISO/IEC 27001Hosting provider
  • BSI C5Hosting provider

Want Haidy® at your institution too?

Klara is the speech recognition inside Haidy®. We'll show you how it works with your reports and your workflows.

Klara listens, Haidy® turns it into