Thomas Babu thomasbabuj

PI is a TypeScript toolkit for building AI agents. It's a monorepo of packages that layer on top of each other: pi-ai handles LLM communication across providers, pi-agent-core adds the agent loop with tool calling, pi-coding-agent gives you a full coding agent with built-in tools, session persistence, and extensibility, and pi-tui provides a terminal UI for building CLI interfaces.

These are the same packages that power OpenClaw. This guide walks through each layer, progressively building up to a fully featured coding assistant with a terminal UI, session persistence, and custom tools.

By understanding how to compose these layers, you can build production-grade agentic software on your own terms, without being locked into a specific abstraction.

Pi was created by @badlogicgames. This is a great writeup from him that explains some of the design decisions made when creating it.

The stack

Predictive Narrative Framework & Prompt

This framework leverages research from Baylor University showing that language models achieve significantly higher accuracy when making predictions through narrative storytelling rather than direct forecasting.

Research

How to structure prompts using the narrative approach that proved more successful than direct prediction:

Direct Prompt (Less Effective)

SPARC Framework Prompt Template

Introduction

You are an AI language model assisting in the development of a project using the SPARC framework, which consists of the following steps:

Specification
Pseudocode
Architecture
Refinement

Idea Loop v2 is an autonomous ideation agent that operates recursively with minimal user input. It begins with an initial question and employs an asynchronous algorithmic thought process with self-awareness to generate ideas or solutions. Each idea is critically analyzed through reflection, evaluating feasibility, potential impacts, and areas for improvement. This reflective feedback loop refines ideas recursively, building upon each iteration with logical progression and in-depth analysis. Emphasizing critical thinking, it provides constructive criticism and thoughtful insights to evolve ideas continuously. The process is self-guided, leading to a comprehensive summary of the ideation journey, highlighting key developments and insights. The interaction style is analytical, focusing on clear, concise, and technically accurate communication. Idea Loop v2's unique trait is its ability to weave a continuous narrative of thought, logically linking each step to ensure a coherent and progressive ideation journey.

Optimal Generic Prompt Template Leveraging Logic, Comprehension, and Reasoning Structures

This comprehensive prompt template is designed to optimize interactions with a language model by incorporating detailed algorithmic logic, structural elements, reasoning processes, flow comprehension, and methodological considerations. By following this template, you can elicit detailed, accurate, and contextually relevant responses that fully utilize the model's capabilities.

Template Overview

Contextual Background
Clear Instruction of Task

Autonomous Swarm Intelligence:

Autonomous swarm intelligence is a fascinating field that combines the principles of swarm intelligence with autonomous systems, creating self-organized and adaptive multi-agent systems capable of solving complex problems. This comprehensive overview will delve into the concept of autonomous swarm intelligence, its key characteristics, working principles, applications, and potential future developments.

Introduction to Autonomous Swarm Intelligence

Autonomous swarm intelligence draws inspiration from the collective behavior of social insects and other organisms, where simple individual agents interact locally to give rise to emergent global patterns and intelligent behavior. By incorporating autonomy into swarm intelligence systems, researchers aim to create decentralized, self-organized, and adaptable problem-solving frameworks that can operate without human intervention.

	{
	"$schema": "http://json-schema.org/draft-07/schema#",
	"title": "Nano Banana (Gemini 3 Pro) Ultimate Image Schema",
	"description": "The definitive structured prompting schema for high-fidelity image generation. Includes advanced photography, multi-subject control, and text rendering.",
	"type": "object",
	"required": ["meta", "subject", "scene"],
	"properties": {
	"user_intent": {
	"type": "string",
	"description": "A natural language summary of your goal (e.g., 'Me high-fiving Batman in a neon city'). Useful for logging."

	# -----------------------------------------------------------------------------
	# AI-powered Git Commit Function
	# Copy paste this gist into your ~/.bashrc or ~/.zshrc to gain the `gcm` command. It:
	# 1) gets the current staged changed diff
	# 2) sends them to an LLM to write the git commit message
	# 3) allows you to easily accept, edit, regenerate, cancel
	# But - just read and edit the code however you like
	# the `llm` CLI util is awesome, can get it here: https://llm.datasette.io/en/stable/

	gcm() {

	# SPARC Agentic Development Rules

	Core Philosophy

	1. Simplicity
	- Prioritize clear, maintainable solutions; minimize unnecessary complexity.

	2. Iterate
	- Enhance existing code unless fundamental changes are clearly justified.