Requirements
Requirements
Desktop (personal / small team use)
| Item | Minimum |
|---|---|
| OS | Windows 10+, macOS 12+, or Linux (Ubuntu 20.04+/Debian 11+/Fedora 35+) |
| RAM | 4 GB (8 GB recommended) |
| Disk | 2 GB free for the app + space for your local LanceDB |
| Internet | Needed for LLM/embeddings/web search API calls |
Server (VPS / shared instance)
| Item | Minimum | Recommended |
|---|---|---|
| CPU | 2 vCPU | 4+ vCPU |
| RAM | 4 GB | 8 GB+ |
| Disk | 20 GB | 50 GB SSD |
| OS | 64-bit Linux | Ubuntu 24.04 |
Complete mode (Postgres + Qdrant + Redis) adds those three services’ requirements on top — size according to expected data volume.
Network
- Configurable HTTP port (default
8080) — the web chat and its internal endpoints share the same port. - TLS isn’t terminated by Vectora itself — use a reverse proxy (Nginx, Caddy, Traefik) in front in production.
- No additional TCP port is needed — the desktop talks to the local backend via IPC (named pipe/unix socket), not TCP.