Papyr documentation
Understand it
- How it works: Yjs on the clients, a blind relay on the server, rooms, secrets and storage.
- Privacy: what is encrypted, what the server still sees, what Papyr cannot protect.
- Performance: what the server costs, latency under load, what the page costs the browser.
Run it
- Self-hosting: start here. Docker and Compose, a reverse proxy, TLS, a public instance, upgrading, coming from Etherpad.
- Operations: the binary and the container, limits per source, the data directory, expiry, backup and restore, releasing.
- Configuration: every setting, and how to size memory and disk from the limits.
- The admin API: status, rooms, metrics, backup (
openapi.json). - Telemetry: every span, metric and log event.
- The command line: read and edit a room from a terminal or a script; for agents.
- Release notes: what each release changed, and what to change when you upgrade.
Build on it
- Hosting the relay in your own app: three calls put the relay into an ASP.NET Core app, behind your own sign-in (example).
- The relay protocol: the wire contract for writing a client, with test vectors.
The design
spec/ holds the specifications the code follows: the
architecture and its glossary, the server, the
room files, the web clients and the
requirements. They are written for contributors.