AI Provider Setup
Configure optional AI features for enhanced functionality.
AI is Optional
Neural Commander's core features work without any AI provider. The daemon, project audits, session persistence, and desktop dashboard all function out of the box.
AI features like nc chat, smart suggestions, and
natural language queries require either local Ollama or a cloud API key.
Provider Options
| Provider | Type | Cost | Privacy |
|---|---|---|---|
| Ollama Recommended | Local | Free (your hardware) | Maximum - data stays local |
| OpenAI | Cloud | Pay-per-use | Data sent to OpenAI |
| Anthropic | Cloud | Pay-per-use | Data sent to Anthropic |
| Azure OpenAI | Cloud | Pay-per-use | Enterprise compliance |
Option A: Ollama (Recommended)
Ollama runs AI models locally on your machine. Free, private, and NC auto-detects it.
1. Install Ollama
2. Pull a Model
3. NC Auto-Detects
No configuration needed! NC automatically checks for Ollama on startup.
Recommended Models
| Model | Size | Best For |
|---|---|---|
| gemma2:2b | 1.6 GB | Quick responses, low resources |
| llama3.2:3b | 2.0 GB | Balanced performance |
| codellama:7b | 3.8 GB | Code-focused tasks |
| llama3.1:8b | 4.7 GB | Best quality for most tasks |
| qwen2.5-coder:7b | 4.7 GB | Excellent for coding |
Option B: Cloud AI (BYOK)
Bring your own API keys from OpenAI, Anthropic, or other providers. You pay the provider directly - we don't charge for AI access.
OpenAI
Anthropic (Claude)
BYOK Model: We don't markup API costs. You pay your provider directly at their standard rates.
What Works Without AI?
Works Without AI
- ✓ NC Daemon (monitoring)
- ✓ Session Persistence
- ✓ Project Health Audits
- ✓ NC Desktop Dashboard
- ✓ Documentation Scanner
- ✓ Requirements YAML
- ✓ Git Integration
- ✓ Active Alerts
Requires AI Provider
- +
nc chat(interactive AI) - + Pattern Extraction
- + Smart Suggestions
- + Natural Language Queries
- + AI-Enhanced Context
- + Code Analysis
Troubleshooting
Ollama not detected?
Make sure Ollama is running:
Model too slow?
Try a smaller model like gemma2:2b (1.6 GB)
or ensure your machine has enough RAM (8GB+ recommended for 7B models).
Cloud API errors?
Verify your API key is correct and has sufficient credits. Check provider status pages for any outages.
FAQ
Which provider do you recommend?
For privacy, Ollama. For maximum capability, Anthropic Claude or OpenAI GPT-4. Many users start with Ollama and add cloud for complex tasks.
Can I switch providers later?
Yes, anytime. Just update your configuration with nc config set ai.provider [provider].
Do you charge for AI usage?
No. We use a BYOK (Bring Your Own Key) model. Ollama is free (uses your hardware). Cloud providers bill you directly. We don't markup AI costs.
What if I don't configure any AI?
NC works fine without AI. Core features like daemon monitoring, session persistence, project audits, and the desktop dashboard all function normally.