Technical Privacy Architecture & Security Specifications
Most communication platforms rely on "Policy-Based Privacy"—promises that data will be handled carefully or deleted upon request. Our infrastructure is built on Architecture-Based Privacy. We employ a "Privacy by Design" framework, where the system is engineered to avoid the creation of sensitive data in the first place.
By implementing a strategy of aggressive data minimisation, we ensure that our server is not a vault for your data, but a transient relay. We have replaced the standard Matrix persistence model with a hardened, ephemeral stack designed to leave no digital footprint.
The first line of defence is the total elimination of server-side readability.
Encryption protects the content, but our erasure pipeline eliminates the existence of the communication through a multi-stage process.
To maintain a minimal data footprint, the system executes redaction sweeps every 30 minutes. Any data that has reached its 12-hour Time To Live (TTL) is automatically scrubbed from the database. This includes:
While redaction removes the content, our retention policy performs a "hard delete" of the event record itself. These purges are executed in hourly sweeps, ensuring that once data reaches its 36-hour TTL, all remaining metadata and server-side pointers are permanently erased.
To prevent forensic data remanence—where "deleted" data can still be recovered from the physical disk—we execute a FULL VACUUM on our PostgreSQL database every two weeks. This physically rewrites the database files on disk, ensuring that dead tuples are permanently overwritten and physically erased from the storage media.
Beyond the data pipeline, the environment hosting the service is hardened to prevent leakage at the OS and hardware levels.
We minimise the potential attack surface by restricting how accounts are accessed and recovered.
We believe that privacy should be verifiable. Users can verify the integrity of our system through two distinct channels: