How does your app recognize chemical formulas and specialized terminology? With other apps, I constantly had to replay research recordings and correct nonsensical text and hallucinations.

We offer two hybrid models: the Standard Sync Model (OpenAI Whisper v3 + Deepgram Nova-2) and the Premium Sync Model (ElevenLabs Scribe v2 + Deepgram Nova-3). Our engineers developed an AI architecture in which models from different providers complement and enhance one another. We also apply multiple stages of filtering to both the incoming audio and the text returned by the models. These solutions were developed entirely by RiverScript and significantly improve the quality of the final transcripts. Both of our transcription models have performed exceptionally well when transcribing complex, specialized terminology and formulas. However, we recommend the Standard Sync Model (available on the Free Start and Pro plans) for most everyday needs, including studying. We recommend the Premium Sync Model (available on the Premium plan) for enterprise-grade transcription when the highest level of accuracy is essential.

If a respondent has an accent, can RiverScript convert their speech into clear text?

Yes. Even RiverScript’s Standard Sync Model (OpenAI Whisper v3 + Deepgram Nova-2) is sufficient for this task. However, if the respondent’s accent is so strong that you can barely understand them, try the Premium Sync Model (ElevenLabs Scribe v2 + Deepgram Nova-3). It delivers exceptionally clean results using the latest transcription technology and converts speech into text while completely eliminating incorrect word recognition.

Can RiverScript translate qualitative research transcripts? Is it like Google Translate?

RiverScript translates qualitative research transcripts and other types of transcripts into 100 languages. The full list is available in the documentation. RiverScript uses an OpenAI model for translation, so the quality is similar to what you would get with ChatGPT. However, there is no chat length limit, so you can translate the transcript of even an eight-hour research recording into another language or several languages at once while maintaining the same translation quality. It is much better than Google Translate.

What else can RiverScript do besides transcribing qualitative research recordings?

RiverScript can add correct punctuation, improve grammar, and organize transcripts into meaningful paragraphs, while basic AI speech recognition tools usually do not include all of these features. RiverScript can also identify and distinguish between speakers, producing readable dialogue with speaker labels rather than just plain text. It can add timestamps to transcripts and split the text into segments of a suitable length for subtitles. RiverScript can translate transcripts from recorded or uploaded audio and video into more than 100 languages using an OpenAI model, so you get ChatGPT-level translation quality without the length limits of AI chats. This means you can translate even the transcript of a video that is several hours long. The workflow also includes an AI chat feature that lets you ask questions about specific information in the audio or request tasks such as creating a blog post based on your research. All of this is achieved by combining the strengths of several AI models that use the most advanced technologies on the market. It is also made possible by RiverScript's own technological solutions, which make AI transcription and content analysis convenient and easy even for beginners. In addition, RiverScript is currently the only solution on the transcription market that lets you record system audio on Windows and Mac, transcribe it, and translate the resulting transcript all in one place without using bots or third-party apps.

Does RiverScript use research recordings to train AI?

No. RiverScript does not use research recordings to train AI and does not train AI models at all. We use AI models that have already been trained for specific tasks. Our policy is that your data belongs only to you. We never use it to train AI models or share it with third-party companies for that purpose.

I want to transcribe a focus group video that may contain sensitive information. How does RiverScript store my data?

RiverScript is hosted in Hetzner Cloud data centers in Helsinki, Finland, within the European Union. Detailed information about recording and transcription security is available in the Security section of our Docs. We retain audio recordings for 7 days for your convenience, while transcripts, translations, and AI responses are stored until you choose to delete them.

Verbatim AI transcription of focus groups with multi-speaker labels for qualitative research

RiverScript handles the routine work while you analyze the data

RiverScript is European AI transcription software from Finland that converts focus group research into verbatim word-for-word text transcript. It transcribes audio and video recordings of group interviews in hundreds of file formats, up to 50 GB and 8 hours long. RiverScript also supports bulk transcription, so you can transcribe dozens of recordings at once.

Multi-speaker transcription

The app turns your recordings into publication-ready material: accurate text with correct grammar and punctuation. It can also identify what different respondents said and label their remarks automatically. The app returns both formatted text suitable for publication and subtitles with timecodes, if you select that option in the menu before transcribing. RiverScript also supports speaker diarization, which you can enable in the menu.

*Please note that the Detect Speakers feature is available only to Premium plan users.

No false matches

RiverScript AI Sync Models recognize complex terminology and formulas without false matches or loss of meaning by using algorithms that filter the audio before transcription and the transcript after the AI model returns it. This makes routine work easier for researchers studying medicine, the exact sciences, law, and business. You can also transcribe not only focus group interviews, but individual in-depth interviews as well.

The software is available for both Mac and Windows on the download page.

AI transcription and research ethics

When transcription is complete, you can rename each focus group participant with one click. For example, you can name them Moderator, Respondent 4 (Male), Respondent 8, and so on. To do this, click a label from a particular respondent, then give that respondent a new name. All their other labels will be renamed automatically. This lets you anonymize your research participants.

If you need a legal document to have your research application approved, you can get one to submit to your university here: RiverScript Research Ethics.

AI Research Assistant

With the RiverScript AI feature Summarize & Ask, you can ask AI Research Assistant a question, for example, to quickly check a hypothesis: “Did any of the experts mention issue X? If so, in what context?” You can also get a brief summary of an entire qualitative research lasting many hours with one click, along with clickable key moments for listening to specific parts of the audio.

Searching for quotes by keyword

If you need to find every mention of a word, click anywhere in the text and press Ctrl+F on Windows or Command ⌘+F on Mac. All matching words will then be highlighted in the text.

Export to your ecosystem

You can download the transcript in various text formats, including PDF and DOCX, or copy the text as Markdown into software for writing dissertations and PhD theses as an Appendix chapter, then format the transcript according to your university’s academic standards.

Please note that no AI can replace a researcher’s observations of group dynamics. RiverScript cannot act as an eyewitness recording the behavior of focus group participants, nor does it identify their individual characteristics or emotions.

If you first want to learn how to install the RiverScript app and record system audio for transcription, we recommend reading the separate guides for Windows and Mac.