Privacy Policy
Effective Date: October 2026 • ORATOR Speech Intelligence Platform Prototype
1. Handling of Uploaded Audio Files
When you upload audio files to ORATOR for acoustic or linguistic evaluation, the audio is processed locally within the running instance environment. In the default autonomous demo configuration, speech audio is stored locally in temporary storage for waveform rendering and signal analysis. Uploaded audio files are not repurposed, redistributed, or sold.
2. Local Database Architecture
ORATOR uses a local SQLite database file (dev.db) managed via Prisma. All analysis records, extracted timestamps, rubric evaluations, and user-generated sessions remain strictly local to your instance unless you configure an external hosted database.
3. Third-Party AI Services
If you explicitly provide external API credentials (such as Google Gemini or OpenAI Whisper keys), your transcripts and audio segments are transmitted to those respective third-party service providers solely to perform inference. If no API keys are provided, ORATOR utilizes internal deterministic signal extraction algorithms, and zero data leaves your local machine.
4. User Control & Data Retention
Users retain complete ownership and control over their uploaded speeches, custom transcripts, and evaluation outputs. Speeches can be deleted from the library at any time via the user interface or database pruning.
5. Prototype Notice
This application is an engineering prototype developed for Hackathon Track C (Contrastive Speech Analytics & Temporal Flaw Grounding). We make no representations regarding formal enterprise certifications (such as HIPAA, SOC2, or GDPR) beyond the described local operational architecture.