System Requirements¶
Minimum Requirements¶
Server (Production)¶
| Component | Minimum | Recommended |
|---|---|---|
| CPU | 4 cores | 8+ cores |
| RAM | 8 GB | 16+ GB |
| Storage | 50 GB SSD | 100+ GB SSD |
| OS | Ubuntu 22.04 LTS / RHEL 8+ | Ubuntu 24.04 LTS |
| Network | 100 Mbps | 1 Gbps |
Storage Depends on Usage
Storage requirements depend on the volume of files in Knowledge Bases and conversation history. Plan for capacity growth as data accumulates.
Server (with Local LLM — Ollama)¶
Running AI Models locally (Ollama) requires additional resources:
| Component | Minimum (7B Model) | Recommended (70B Model) |
|---|---|---|
| CPU | 8 cores | 16+ cores |
| RAM | 16 GB | 64+ GB |
| GPU | NVIDIA GPU 8 GB VRAM | NVIDIA GPU 24+ GB VRAM |
| Storage | 100 GB SSD | 500+ GB NVMe SSD |
GPU for Local LLM
- Primarily supports NVIDIA CUDA (requires NVIDIA Driver + CUDA Toolkit)
- Apple Silicon (M1/M2/M3/M4) supported via Ollama on macOS
- AMD ROCm supported on select models
Software Requirements¶
Required Software¶
| Software | Version | Purpose |
|---|---|---|
| Docker | 24.0+ | Container runtime |
| Docker Compose | 2.20+ | Container orchestration |
| PostgreSQL + pgvector | 15+ / 0.7+ | Database + vector search |
| Git | 2.30+ | Source code management |
Optional Software¶
| Software | Version | Purpose |
|---|---|---|
| Nginx | 1.24+ | Reverse proxy / SSL termination |
| NVIDIA Container Toolkit | Latest | GPU support for Ollama |
| Certbot | Latest | SSL certificate (Let's Encrypt) |
Client Requirements (Web Browser)¶
| Browser | Minimum Version |
|---|---|
| Google Chrome | 100+ |
| Microsoft Edge | 100+ |
| Mozilla Firefox | 100+ |
| Safari | 16+ |
Browser Features Required
- JavaScript enabled
- WebSocket support
- Web Audio API (for audio features)
- Clipboard API (for copy/paste)
Network Requirements¶
Ports¶
| Port | Service | Direction |
|---|---|---|
| 443 (HTTPS) | Web Application | Inbound |
| 80 (HTTP) | Redirect to HTTPS | Inbound |
| 8080 | Backend API (internal) | Internal only |
| 5180 | Frontend dev server (internal) | Internal only |
| 11434 | Ollama API (if used) | Internal only |
External Connectivity¶
When using Cloud AI Providers, the system must be able to reach the following external endpoints:
| Service | Endpoint | Port |
|---|---|---|
| OpenAI | api.openai.com |
443 |
| Anthropic | api.anthropic.com |
443 |
| Google AI | generativelanguage.googleapis.com |
443 |
| OpenRouter | openrouter.ai |
443 |
Firewall / Proxy Configuration
If your organization uses a firewall or proxy, you must whitelist the endpoints of the AI Providers you intend to use. The application does not support HTTP Proxy internally — configure the proxy at the OS or Docker level.
Sizing Guide¶
By Number of Users¶
| Size | Concurrent Users | CPU | RAM | Storage |
|---|---|---|---|---|
| Small | 1–50 | 4 cores | 8 GB | 50 GB |
| Medium | 50–200 | 8 cores | 16 GB | 200 GB |
| Large | 200–1,000 | 16 cores | 32 GB | 500 GB |
| Enterprise | 1,000+ | 32+ cores | 64+ GB | 1 TB+ |
Cloud AI vs Local LLM
The table above applies to Cloud AI Provider usage only. If using Local LLM, add resources as specified in the Local LLM section above.