Privacy Policy
Last updated: April 27, 2026 · Version 1.0
Who is responsible
AudioMap is a brand operated by José Díaz from Madrid, Spain. For any data privacy questions, write to [email protected].
If you reside in the European Union, your data is processed under the General Data Protection Regulation (GDPR) and Spanish Law 3/2018 (LOPDGDD).
What data we collect
- Account and profile: email, name, and identity provider (Keycloak federated at auth.codelabs.studio).
- Audio and transcriptions: files you upload and the text we generate from them. Accessible only to you.
- Consents: we log every permission you accept with timestamp, IP, user-agent, and version of the legal text accepted. This table is immutable and reconstructible at any historical point.
- Usage: minutes consumed, AI models used, tokens. Used for quota enforcement and service improvement. Never sold.
How we use your data
- Transcribe your audio and generate AI Notes and chat responses.
- Enforce your plan (available minutes, contracted storage).
- Support and communication (only from @audiomap.com addresses).
- Legal compliance and response to founded court orders.
Your audio is never used to train AI models. Not ours, not third parties'. This is a contractual obligation, not an optional preference.
Subprocessors that touch your data
The following providers process your data exclusively as our subprocessors:
- AssemblyAI (USA) — audio transcription. Data processing agreement signed. Audio is deleted from their system once transcription is complete.
- Anthropic (USA) — AI Notes and chat generation. Anthropic does not train models with data sent via API. Agreement signed.
- OpenAI (USA) — semantic embeddings for RAG chat. OpenAI has not trained on API data since March 2023 without explicit opt-in. Agreement signed.
- Cloudflare R2 — audio storage in a Western Europe bucket. SSE-S3 encryption at rest.
- Dedicated server in Europe — PostgreSQL database, semantic embeddings (local bge-m3 engine), job queue and workers. Under our exclusive management. Transcripts never leave this server to generate vectors.
- Sentry (USA) — error monitoring. Active filters to prevent sending sensitive content (transcripts, audio).
How long we keep your data
- Original audio: while your account is active and you have given explicit consent. If revoked, we physically delete it from R2 within 7 days.
- Transcriptions and AI Notes: while your account is active or until you delete them.
- Consent audit log: 6 years after account closure (minimum legal requirement in Spain for contractual records). This is the only table that survives account closure.
You can exercise your rights of access, rectification, erasure, portability, objection, and restriction by writing to [email protected]. We respond within the legal one-month deadline.
No spy bot, no third-party recording
AudioMap processes files that you upload. We do NOT have a bot that joins your meetings to record your guests without explicit consent. Active class action lawsuits against Otter.ai (Brewer v. Otter.ai, Aug 2025) and Fireflies.ai (Cruz v. Fireflies, Dec 2025) demonstrate that the "self-invited bot" model carries structural legal risk. If you want to record a multi-person meeting, do so explicitly with recorded verbal consent from all participants.