Start your Claude Code session with this approach - no configuration needed:
"Get the current working directory with pwd, export MCP_CLIENT_CWD to that value, then run a health check to verify the correct project context"
Or for indexing:
"First run pwd and export MCP_CLIENT_CWD=$(pwd), then reindex this directory and give me an overview of the project structure"
This natural language approach:
- Works immediately without any configuration changes
- Ensures the MCP server knows your actual working directory
- Can be used in any project without setup
If you're already in the correct directory:
"Reindex this directory, then search for the main files and give me an overview of the project structure"
This will:
- Clear any stale data
- Index all current files
- Search and summarize key components
- Give you a project overview
Add to your shell config (~/.bashrc or ~/.zshrc):
# Auto-index every Claude session
export QDRANT_RAG_AUTO_INDEX=true
alias claude-dev='QDRANT_RAG_AUTO_INDEX=true claude'# From any project directory
~/path/to/qdrant-rag/scripts/start-session.shAfter starting Claude Code, use these commands for instant context:
-
Set Working Directory + Health Check (Recommended first command):
"Run pwd to get current directory, export MCP_CLIENT_CWD=$(pwd), then run a health check to confirm project context" -
Full Reindex + Overview:
"Export MCP_CLIENT_CWD=$(pwd), then reindex the current directory and tell me about the project structure, main files, and recent changes" -
Quick Context Verification:
"Run a health check and show me the current project context" -
Find Key Files:
"Search for README, main, index, app, and config files to understand the project" -
Understand Dependencies:
"Search for package.json, requirements.txt, pyproject.toml, or go.mod and summarize the dependencies"
Add to your project's README or documentation:
# Quick start for this project
alias claude-this='cd /path/to/project && QDRANT_RAG_AUTO_INDEX=true claude'Create .envrc in your project (using direnv):
# .envrc
export QDRANT_RAG_AUTO_INDEX=true
export QDRANT_RAG_DEBOUNCE=2.0Start with recent changes context:
"Reindex this directory, then search for files modified in the last week and summarize recent changes"
For large projects, index strategically:
"First index just the src directory, then search for the main entry points"
"Index all Python files and search for class definitions"
"Index configuration files and tell me about the project setup"
Always verify your context is loaded:
"What files are currently indexed for this project?"
"Show me the health check status"
"Search for a known file to verify indexing worked"