Transcription
Transparency is central to us at Intric. When you use the Transcription feature to process and summarize recorded audio files, specific processes are in place to ensure your privacy and secure data handling.
Flow overview
The process from your audio upload to a finished text transcription and summary occurs through a secure, SSL/TLS encrypted interaction between the Intric platform (where your data is managed) and your selected LLM provider (for example, Berget AI).
Step-by-step: How your data is handled
Step 1 — User
You upload an audio file.
Input: You upload a recorded audio file directly into the Intric application.
Step 2 — Intric (Server - Initial Processing)
Intric receives and secures the file.
Action: The server receives the audio file and loads it into RAM. It is immediately encrypted and stored securely on disk within Intric’s database. When ready, the audio file is loaded for processing. Before sending the audio to the AI model, Intric strictly removes all metadata—meaning no information about who uploaded it or which organization it belongs to is ever attached.
Step 3 — Selected assistant model (LLM Provider)
Transcription and Summary Processing.
Input: The LLM provider receives only the audio data. Action: The AI model processes the audio data to generate a text transcription, and subsequently processes it to generate a summary. It has zero context about the identity of the user. Immediately after the transcription and summary are sent back, the LLM provider deletes the audio input and the generated responses.
Step 4 — Intric (Server - Result Handling)
Intric saves the results.
Action: Intric securely receives the processed text. The transcription text and the summary are both saved directly to Intric’s secure database.
Step 5 — User
You receive the final result.
Action: Intric loads the saved transcript and summary and presents the final results to you in the user interface.
Data sharing and privacy
To protect your and your organization’s privacy, we apply the principle of data minimization. All transmissions between Intric and the LLM provider are protected via SSL/TLS encryption. The LLM provider only gets access to the audio content absolutely necessary to perform the task—no user identity ever leaves your infrastructure.
Furthermore, we have strict zero data retention clauses in all our contracts with LLM providers. This guarantees that your audio files and the generated texts are never saved by the provider after the request is fulfilled, nor are they ever used to train their AI models.
In the table below, you can see exactly what data is sent to the LLM and what is kept completely private.
| This is sent to the LLM | This is NOT sent to the LLM |
|---|---|
| The recorded audio file / audio data | User ID, username, or name |
| System instructions for transcription and summarization | Email addresses or IP addresses |
| Authentication tokens | |
| Database IDs |
Summary
When you use the Transcription feature in Intric, your files and identity remain completely private. Your audio is encrypted and stored locally on Intric’s servers. The chosen LLM provider only ever sees the isolated audio data required to transcribe and summarize the recording. Thanks to secure SSL/TLS transmission and zero data retention agreements, your audio and text data are immediately forgotten by the external AI model once the results are delivered.