Fixing Ollama Timeouts

Why Does DevDiff Time Out?

Ollama processes AI requests on your machine. Large diffs (many files, many lines) need more processing time. A 30-second timeout may be too short for 70+ files.

Quick Fix

# Use minimal depth for faster results
devdiff generate --depth minimal

# Or process fewer files at once
git add src/auth/     # Stage only one directory
devdiff generate
git add src/api/      # Then stage the next
devdiff generate

Permanent Fixes

1. Increase Timeout

// .devdiff.config.js
export default {
  ai: {
    ollama: {
      timeout: 120000, // 2 minutes (in milliseconds)
    },
  },
};

2. Use a Faster Model

# Smaller models are faster (but less detailed)
ollama pull llama3.2:1b  # Very fast, basic analysis

# Update config
# .devdiff.config.js
export default {
  ai: {
    providers: [
      { url: 'ollama://llama3.2:1b', priority: 1 }
    ]
  }
}

3. Enable Chunking

DevDiff v1.0.5+ automatically splits large diffs into chunks. No configuration needed — it detects when chunking is necessary.

4. Check Your Hardware

Problem	Symptom	Fix
Low RAM	System slows during AI	Close other apps, use smaller model
Slow CPU	AI takes >2 min	Use cloud AI (OpenAI, Anthropic)
Disk I/O	Model loading is slow	Move models to SSD

Still Timing Out?

1. Run devdiff doctor for diagnostics
2. Check Ollama logs: ollama logs
3. Try a simple test: ollama run llama3.2:3b "hello"
4. If simple test works but DevDiff doesn't, report an issue