Grepai Troubleshooting

GrepAI Troubleshooting

This skill provides solutions for common GrepAI issues and diagnostic procedures.

When to Use This Skill

• GrepAI not working as expected
• Search returning poor results
• Index not updating
• Connection or configuration errors

Quick Diagnostics

Run these commands to understand your setup:

# Check GrepAI versiongrepai version
# Check project statusgrepai status
# Check Ollama (if using)curl http://localhost:11434/api/tags
# Check configcat .grepai/config.yaml

Common Issues

Issue: "Index not found"

Symptom:

Error: Index not found. Run 'grepai watch' first.

Cause: No index has been created for this project.

Solution:

# Initialize if neededgrepai init
# Create the indexgrepai watch

Issue: "Cannot connect to embedding provider"

Symptom:

Error: Cannot connect to Ollama at http://localhost:11434

Causes:

1. Ollama not running
2. Wrong endpoint configured
3. Firewall blocking connection

Solutions:

1. Start Ollama:

ollama serve

2. Check endpoint in config:

embedder:  endpoint: http://localhost:11434  # Verify this

3. Test connection:

curl http://localhost:11434/api/tags

Issue: "Model not found"

Symptom:

Error: Model 'nomic-embed-text' not found

Cause: The embedding model hasn't been downloaded.

Solution:

# Download the modelollama pull nomic-embed-text
# Verifyollama list

Issue: Search returns no results

Symptom: Searches return empty or very few results.

Causes:

1. Index is empty
2. Files are being ignored
3. Query too specific

Solutions:

1. Check index status:

grepai status# Should show files > 0 and chunks > 0

2. Verify files are being indexed:

# Check ignore patterns in configcat .grepai/config.yaml | grep -A 20 "ignore:"

3. Try broader query:

grepai search "function"  # Very broad test

Issue: Search returns irrelevant results

Symptom: Results don't match what you're looking for.

Causes:

1. Query too vague
2. Boosting not configured
3. Wrong content indexed

Solutions:

1. Improve query (see grepai-search-tips skill):

# Badgrepai search "auth"
# Goodgrepai search "user authentication middleware"

2. Configure boosting to penalize tests:

search:  boost:    enabled: true    penalties:      - pattern: /tests/        factor: 0.5

3. Check what's indexed:

grepai status

Issue: Index is outdated

Symptom: Recent file changes aren't appearing in search results.

Causes:

1. Watch daemon not running
2. Debounce delay
3. File not in indexed extensions

Solutions:

1. Check daemon status:

grepai watch --status

2. Restart daemon:

grepai watch --stopgrepai watch --background

3. Force re-index:

rm .grepai/index.gobgrepai watch

Issue: "Config not found"

Symptom:

Error: Config file not found at .grepai/config.yaml

Cause: GrepAI not initialized in this directory.

Solution:

grepai init

Issue: Slow indexing

Symptom: Initial indexing takes very long.

Causes:

1. Large codebase
2. Slow embedding provider
3. Not enough ignore patterns

Solutions:

1. Add ignore patterns:

ignore:  - node_modules  - vendor  - dist  - build  - "*.min.js"

2. Use faster model:

embedder:  model: nomic-embed-text  # Smaller, faster

3. Use OpenAI for speed (if privacy allows):

embedder:  provider: openai  model: text-embedding-3-small  parallelism: 8

Issue: Slow searches

Symptom: Search queries take several seconds.

Causes:

1. Very large index
2. GOB storage on large codebase
3. Embedding provider slow

Solutions:

1. Check index size:

ls -lh .grepai/index.gob

2. For large indices, use Qdrant:

store:  backend: qdrant

3. Limit results:

grepai search "query" --limit 5

Issue: Trace not finding symbols

Symptom: grepai trace callers returns no results.

Causes:

1. Function name spelled wrong
2. Language not enabled for trace
3. Symbols index out of date

Solutions:

1. Check exact function name (case-sensitive)

2. Enable language in config:

trace:  enabled_languages:    - .go    - .js    - .ts

3. Re-build symbol index:

rm .grepai/symbols.gobgrepai watch

Issue: MCP not working

Symptom: AI assistant can't use GrepAI tools.

Causes:

1. MCP config incorrect
2. GrepAI not in PATH
3. Working directory wrong

Solutions:

1. Test MCP server manually:

grepai mcp-serve

2. Check GrepAI is in PATH:

which grepai

3. Verify MCP config:

# Claude Codecat ~/.claude/mcp.json
# Cursorcat .cursor/mcp.json

Issue: Out of memory

Symptom: GrepAI crashes or system becomes slow.

Causes:

1. Large embedding model
2. Very large index in GOB format
3. Too many parallel requests

Solutions:

1. Use smaller model:

embedder:  model: nomic-embed-text  # Smaller

2. Use PostgreSQL or Qdrant instead of GOB

3. Reduce parallelism:

embedder:  parallelism: 2

Issue: API key errors (OpenAI)

Symptom:

Error: 401 Unauthorized - Invalid API key

Solutions:

1. Check environment variable:

echo $OPENAI_API_KEY

2. Ensure variable is exported:

export OPENAI_API_KEY="sk-..."

3. Check key format in config:

embedder:  api_key: ${OPENAI_API_KEY}  # Uses env var

Diagnostic Commands

Full System Check

#!/bin/bashecho "=== GrepAI Diagnostics ==="
echo -e "\n1. Version:"grepai version
echo -e "\n2. Status:"grepai status
echo -e "\n3. Config:"cat .grepai/config.yaml 2>/dev/null || echo "No config found"
echo -e "\n4. Index files:"ls -la .grepai/ 2>/dev/null || echo "No .grepai directory"
echo -e "\n5. Ollama (if using):"curl -s http://localhost:11434/api/tags | head -5 || echo "Ollama not responding"
echo -e "\n6. Daemon:"grepai watch --status 2>/dev/null || echo "Daemon not running"

Reset Everything

If all else fails, complete reset:

# Remove all GrepAI datarm -rf .grepai
# Re-initializegrepai init
# Start fresh indexgrepai watch

Getting Help

If issues persist:

1. Check GrepAI documentation: https://yoanbernabeu.github.io/grepai/
2. Search issues: https://github.com/yoanbernabeu/grepai/issues
3. Create new issue with:
   • GrepAI version (grepai version)
   • OS and architecture
   • Config file (remove secrets)
   • Error message
   • Steps to reproduce

Output Format

Diagnostic summary:

🔍 GrepAI Diagnostics
Version: 0.24.0Project: /path/to/project
✅ Config: Found (.grepai/config.yaml)✅ Index: 245 files, 1,234 chunks✅ Embedder: Ollama (connected)✅ Daemon: Running (PID 12345)❌ Issue: [Description if any]
Recommended actions:1. [Action item]2. [Action item]