hogli: one CLI for two drivers (our engineers and their agents) | Miami .
Cities (231) ▾

Members-Only

Recent Talks & Demos are for members only

Exclusive feed

You must be an AI Tinkerers active member to view these talks and demos.

Sign in View FAQ
June 25, 2026 · Miami

hogli: CLI for Agent Devboxes

See how a single CLI, hogli, manages complex development stacks for both engineers and AI agents, enabling quick setup and remote debugging.

Python Click MCP Coder Docker Compose
Overview

The PostHog DevEx team has been building out “hogli”, a Python-based CLI that both our engineers and our coding agents drive the same way. PostHog’s dev stack is about 20+ processes and around the same amount of containers. It’s too RAM-hungry to easily run twice on a laptop, so we set up disposable remote boxes (using Coder.com) that hogli can drive and manage. An engineer (or an agent) can get a working stack in minutes, together with a live URL for clicking around. hogli devbox:task hands Claude Code its own box, hogli up -d runs the process manager (phrocs, our custom version of mprocs) detached behind a JSON socket, and a small MCP server on that socket lets the agent read process status, grep logs, and restart processes without a terminal. I’ll demo both drivers live on a running box: the detached stack and its socket, then I break a process and watch the agent read the logs, find the cause, and bring it back.

Links
Tech stack
  • Python
    Python: The high-level, general-purpose language built for readability, powering everything from web backends to advanced machine learning models.
    Python is the high-level, general-purpose language prioritizing clear, readable syntax (via significant indentation), ensuring rapid development for any team . Its ecosystem is massive: use it for robust web development with frameworks like Django and Flask, or leverage its power in data science with libraries such as Pandas and NumPy . The Python Package Index (PyPI) provides thousands of community-contributed modules, offering immediate solutions for tasks from network programming to GUI creation . The language is actively maintained by the Python Software Foundation (PSF), with the stable release currently at Python 3.14.0 (as of November 2025) .
  • Click
    Click is a Python package that simplifies creating composable, production-ready command line interfaces with minimal decorator-based code.
    Writing command line tools in Python often gets bogged down in boilerplate argument parsing. Click solves this by using clean Python decorators (like @click.command and @click.option) to bind CLI logic directly to your functions. It handles the heavy lifting out of the box: it automatically generates clean help pages, enforces strict type validation, and supports nested command groups. Whether you are building a simple utility or a complex developer tool, Click keeps your codebase modular and your terminal interface highly intuitive.
  • MCP
    MCP is the open-source standard for securely connecting AI agents (like LLMs) to external tools, data, and enterprise workflows.
    The Model Context Protocol (MCP) functions as a standardized integration layer: think of it as a USB-C port for AI applications. Developed and open-sourced by Anthropic, this protocol allows large language models (LLMs) to access real-time context and execute actions via external tools like GitHub, Jira, or proprietary databases . It uses a simple JSON-RPC interface to define tools, schemas, and endpoints, which enables AI agents to perform complex, state-changing tasks—such as creating a GitHub issue or running a test script—rather than just generating text . MCP is essential for building agentic AI systems that can autonomously pursue goals and operate within defined safety and permission boundaries .
  • Coder
    Coder is an open-source, self-hosted platform that provisions secure cloud development environments and manages AI coding agents on your own infrastructure.
    Coder shifts software development off local machines and into secure, self-hosted cloud infrastructure (CDEs) using Terraform to define workspaces as Kubernetes pods, Docker containers, or virtual machines. By centralizing the development environment, engineering teams onboard in seconds, secure their source code, and automatically shut down idle cloud resources to slash compute costs. The platform also provides a governed, self-hosted control plane for AI coding agents: developers run native AI loops directly on company infrastructure without exposing API keys or LLM credentials inside the workspace.
  • Docker Compose
    Define, configure, and run multi-container Docker applications using a single YAML file.
    Docker Compose streamlines multi-container application management: it uses a declarative YAML file (typically `compose.yaml`) to specify all services, networks, and volumes for your application stack. This configuration allows developers to build, start, and stop the entire environment—for instance, a web service linked to a Redis cache and a PostgreSQL database—with one command: `docker compose up`. It’s the essential tool for efficient local development, testing, and CI/CD workflows, ensuring consistency across environments.
Events Gallery FAQ Sponsor