Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Course Outline
Introduction to Pi
- Understanding what Pi is and its core philosophy of minimalism and control
- Comparing Pi with Claude Code, Codex CLI, and Cursor
- An overview of the four built-in tools: read, write, edit, and bash
- Identifying scenarios where Pi is preferred over more complex agents
Installation and Setup
- Installing Pi via npm and various package managers
- Configuring API keys for over 15 supported providers
- Setting up the working directory and context files
- Running Pi in both local and Docker environments
Core Terminal Workflows
- Generating code from simple text prompts
- Precisely reading and editing files
- Safely executing bash commands within the agent
- Managing session state and performing context resets
Building Custom Skills
- Understanding the TypeScript skill API structure
- Writing custom operations for reading, writing, and editing files
- Integrating with bash execution for specialized workflows
- Testing and debugging skills in isolation
Prompt Templates and Themes
- Tailoring system prompts for different languages and frameworks
- Creating reusable prompt templates to ensure team consistency
- Packaging themes as shareable modules
Extending Pi with npm Packages
- Bundling skills as 'pi' packages
- Publishing to npm and Git repositories
- Installing community-contributed packages
- Managing dependency versions and compatibility issues
Provider Flexibility
- Switching between Anthropic, OpenAI, Google, and other backend providers
- Optimizing for cost, speed, and quality on a per-task basis
- Running entirely local models for sensitive code handling
Practical Use Cases
- Automated testing and test generation workflows
- Refactoring legacy code using custom skill chains
- Generating and maintaining documentation
- Integrating with CI/CD pipelines with minimal agent overhead
Security and Control
- Understanding the limitations of the 300-word system prompt boundary
- Requiring manual approval for write operations
- Sandboxing bash execution to protect untrusted environments
- Audit logging all agent actions for transparency
Summary and Next Steps
- Recapping principles of minimalism, control, and extensibility
- Overview of community resources and contribution guidelines
- Exploring advanced topics for experienced Pi users
Requirements
- Experience with terminal-based development and bash scripts
- Fundamental knowledge of TypeScript or JavaScript
- Familiarity with npm package management
Audience
- Developers looking for a lightweight, hackable coding agent
- Senior engineers seeking full control over agent internals
- Teams embracing terminal-first development practices
14 Hours