Troubleshooting¶
Common issues and solutions.
API Key Issues¶
Error: API key not found
# Check if key is set
echo $OPENAI_API_KEY
# Set the key
export OPENAI_API_KEY="sk-..."
Error: Invalid API key
Verify the key is correct (no extra spaces)
Check if the key has been revoked
Regenerate from provider dashboard
Provider Connection¶
Error: Connection refused
# Check internet connectivity
curl https://api.openai.com
# For Ollama, ensure it's running
ollama serve
Error: Rate limit exceeded
Wait and retry
Reduce request frequency
Consider using a different provider
Model Issues¶
Error: Model not found
# List available models
perspt --list-models
# For Ollama
ollama list
Error: Context length exceeded
Use a model with longer context
Reduce conversation history
Clear chat and start fresh
Agent Mode Issues¶
Agent stuck in retry loop
Check LSP is working:
ty check file.py
Lower stability threshold:
perspt agent --stability-threshold 0.5 "task"
Check for unsolvable errors in code
High energy despite clean code
Run tests manually:
pytest -vCheck LSP diagnostics
Adjust energy weights:
perspt agent --energy-weights "0.5,0.5,1.0" "task"
Agent aborted unexpectedly
# Check session status
perspt status
# Resume if possible
perspt resume
TUI Issues¶
Terminal rendering problems
Ensure terminal supports 256 colors
Try a different terminal emulator
Check
$TERMenvironment variable
Keyboard shortcuts not working
Check terminal keybinding conflicts
Try raw mode:
perspt --raw
Configuration Issues¶
Config file not found
# Create config directory
mkdir -p ~/.perspt
# Initialize config
perspt config --edit
Invalid config format
Check TOML syntax:
# Validate TOML
python -c "import tomllib; tomllib.load(open('~/.perspt/config.toml', 'rb'))"
Build Issues¶
Compilation errors
# Update Rust
rustup update
# Clean and rebuild
cargo clean
cargo build --release
Missing dependencies
# macOS
brew install openssl pkg-config
# Ubuntu/Debian
sudo apt install libssl-dev pkg-config
Performance Issues¶
Slow response times
Try Groq for faster inference
Use smaller models (e.g.,
gemini-3-flash)Check network latency
High memory usage
Reduce conversation history
Use streaming mode
Restart application periodically
Debug Mode¶
Enable verbose logging:
perspt -v chat
perspt -v agent "task"
# Maximum verbosity
RUST_LOG=debug perspt chat
Getting Help¶
Check documentation: This book
GitHub Issues: github.com/eonseed/perspt/issues
Logs: Check
~/.perspt/logs/
See Also¶
CLI Reference - Command reference
Configuration - Configuration guide