pocket-joe 0.1.0.4

A correct, simple, performant, and pythonic framework for building durable AI agents

PocketJoe

LLM Agents are just agents...

• Agents are policies
• A policy reasons over observations and chooses a batch of actions
• A policy can be any mix of LLM-based, human-in-the-loop, or heuristic

Semantics

An agent system using Reinforcement Learning theory with LLM semantics as first class

• policy: all code/logic/llm are policies
• observations - the set of observations for the policy to reason over
• options - a set of optional sub policies that the policy can choose
• selected_actions - the set of concurrent actions the policy chose to take
• Message: a shared dataclass for observations and actions that aligns with llm semantics

LLM semantics as platform semantics

In LLM APIs, everything is a Message. We adopt this as our universal unit:

• Input: observations: list[Message] (what the policy sees)
• Output: selected_actions: list[Message] (what the policy does)

Key insight: The runtime automatically invokes all option calls and injects the results back as observations. Your policy just returns requests; the platform handles execution.

Everything is a Policy

An LLM policy that can call other policies:

@policy.tool(description="Calls LLM with tool support")
async def openai_llm_policy_v1(observations: list[Message], options: list[OptionSchema]) -> list[Message]:
    """LLM policy that calls OpenAI GPT-4 with tool support.
    Args:
        observations: List of Messages representing the conversation history + new input
        options: Set of allowed options the LLM can call
    """
        openai = AsyncOpenAI()
        response = await openai.chat.completions.create(
            model="gpt-4",
            messages=observations_to_completions_messages(observations),
            tools=options_to_completions_tools(options))
        return completions_response_to_messages(response)

A simple heuristic policy:

@policy.tool(description="Performs a web search and returns results.")
async def web_seatch_ddgs_policy(query: str,) -> list[Message]:
        """Performs a web search and returns results.        
        Args:
            query: The search query string to search for
        """
        results = DDGS().text(query, max_results=5)
        results_str = "\n\n".join([f"Title: {r['title']}\nURL: {r['href']}\nSnippet: {r['body']}" for r in results])
        return [Message(
            actor="web_seatch_ddgs_policy",
            type="action_result",
            payload={"content": results_str}
        )]

An orchestrator policy that coordinates LLM + search:

@policy.tool(description="Orchestrates LLM with web search tool")
async def search_agent(prompt: str) -> list[Message]:
        """Orchestrator that gives the LLM access to web search.
        Args:
            prompt: The user prompt to process
        """
        ctx = AppContext.get_ctx()
        system_message = Message(actor="system", type="text",
            payload={"content": "You are an AI assistant that can use tools to help answer user questions."})
        prompt_message = Message(actor="user", type="text", payload={"content": prompt})

        history = [system_message, prompt_message]
        while True:
            selected_actions = await ctx.llm(observations=history, options=OptionSchema.from_func([ctx.web_search]))
            history.extend(selected_actions)
            if not any(msg.type == "action_call" for msg in selected_actions):
                break

        return history

Use AppContext for registry (gives IDE type hints):

class AppContext(BaseContext):
    def __init__(self, runner):
        super().__init__(runner)
        self.llm = self._bind(openai_llm_policy_v1)
        self.web_search = self._bind(web_seatch_ddgs_policy)
        self.search_agent = self._bind(search_agent)

Enjoy:

async def main():
    runner = InMemoryRunner()
    ctx = AppContext(runner)
    result = await ctx.search_agent(prompt="What is the latest Python version?")
    print(f"\nFinal Result: {result[-1].payload['content']}")

Why this matters:

• Same interface for LLM, human, heuristic policies
• All policy parameters are optional (define what you need)
• Type-safe composition with IDE support
• Enables evolution: human → heuristic → LLM with no refactoring

A correct, simple, performant, and pythonic framework for building durable AI agents.

"There is no flow, only Policies and Actions."

Getting Started

Prerequisites

• Python 3.12+

Installation

uv add pocket-joe

Or with pip:

pip install pocket-joe

To install with example dependencies:

uv add pocket-joe --extra examples
# or
pip install pocket-joe[examples]

Development Setup

git clone https://github.com/Sohojoe/pocket-joe.git
cd pocket-joe
uv sync --dev --all-extras

Running Examples

Set your API key:

export OPENAI_API_KEY=sk-...

Search Agent (ReAct)

uv run python examples/search_agent.py

YouTube Summarizer

uv run python examples/youtube_summarizer.py

Dev Status

Still in prerelease, things will change

Initial version

• [] Tidy up code - add partly refactored code
• [] Proper tests
• [] Implement more examples from Pocket-Flow

Durable System:

• [] Ledger - Temporal style 'at least once, only one result' replay semantic
• [] Durable Storage wrapper - For long running tasks & replay
• [] Distributed - worker model

Background

Inspired by PocketFlow... I loved PocketFlow but it fell short in a couple of key areas. This is my rewrite that I can actually use.