Ollama Setup

Using Mycelos with Ollama

Run Mycelos completely locally with Ollama — no cloud API keys, no costs, full privacy. Your data never leaves your machine.

Install Ollama

# macOS
brew install ollama

# Linux
curl -fsSL https://ollama.com/install.sh | sh

# Windows
# Download from https://ollama.com/download

Pull a Model

# Recommended for most machines (4GB RAM)
ollama pull gemma3:4b

# For more capable responses (8GB+ RAM)
ollama pull llama3.3

# For coding tasks
ollama pull devstral

Configure Mycelos

During mycelos init, choose Ollama as your provider:

export OLLAMA_API_BASE=http://localhost:11434
mycelos init
# When asked for provider, enter the Ollama URL

Or add Ollama to an existing setup:

mycelos credential store ollama
# Enter: http://localhost:11434

Network Setup (Access from Other Devices)

By default, Ollama only listens on localhost. To use it from another machine (e.g., Mycelos on a Raspberry Pi, Ollama on a Mac):

# On the machine running Ollama:
OLLAMA_HOST=0.0.0.0 ollama serve

To make this permanent:

# macOS (if running as app)
launchctl setenv OLLAMA_HOST 0.0.0.0
# Then restart Ollama app

# macOS / Linux (command line)
echo 'export OLLAMA_HOST=0.0.0.0' >> ~/.zshrc
# Then: source ~/.zshrc && ollama serve

# Linux (systemd service)
sudo systemctl edit ollama
# Add: Environment=OLLAMA_HOST=0.0.0.0
sudo systemctl restart ollama

Then configure Mycelos to point to the Ollama server:

export OLLAMA_API_BASE=http://192.168.1.42:11434  # Ollama machine IP
mycelos serve

Recommended Models

LM Studio (Alternative)

LM Studio provides a GUI for running local models and exposes an OpenAI-compatible API. Mycelos works with LM Studio out of the box:

# LM Studio starts a server at localhost:1234 by default
# Configure as OpenAI provider with local URL:
export OPENAI_API_BASE=http://localhost:1234/v1
mycelos credential store openai
# Enter any string as API key (LM Studio ignores it)

LM Studio advantages:

• Visual model browser and download manager
• GPU acceleration setup is easier (especially on Mac)
• Chat interface for testing models before using with Mycelos

Troubleshooting

”Model too large” / Process killed

Your machine doesn’t have enough RAM. Try a smaller model:

ollama pull gemma2:2b  # Only needs 2GB

“Connection refused”

Ollama isn’t running or is on a different port:

# Check if Ollama is running
curl http://localhost:11434
# Should respond: "Ollama is running"

# If not, start it:
ollama serve

“Slow responses”

Local models are slower than cloud APIs. Tips:

• Use smaller models (2B-4B parameters)
• Enable GPU acceleration (Ollama auto-detects)
• Use an SSD instead of HDD for model storage
• Close other applications to free RAM

Network: “Can’t connect from other device”

# Verify Ollama listens on all interfaces
curl http://YOUR_IP:11434
# If this fails, Ollama is still on localhost only
# Set OLLAMA_HOST=0.0.0.0 and restart