← All projects
Project record02 mentions
Originally shared on Hacker News

Obeya

This app is an open-source spatial workspace for directing AI coding agents like an engineering director. Each project gets a canvas where work is laid out as cards, with an agent running in a git worktree per card. Agent chat is condensed to WhatsApp-style back-and-forth instead of long tool-call logs, and results are typically narrated demo videos that can be commented on by voice to send the agent another round. It supports remote control from a phone and remote workers to scale beyond a single machine.

Last shared
01 / The problem

What it addresses

The author was overwhelmed by walls of text from parallel Claude Code sessions, which limited how many sessions could be steered in parallel to about five.

02 / Source history

Shared over time

Each entry reflects what was shared on that date. Details may have changed since then.

  1. Obeya

    This app is an open-source spatial workspace for directing AI coding agents like an engineering director. Each project gets a canvas where work is laid out as cards, with an agent running in a git worktree per card. Agent chat is condensed to WhatsApp-style back-and-forth instead of long tool-call logs, and results are typically narrated demo videos that can be commented on by voice to send the agent another round. It supports remote control from a phone and remote workers to scale beyond a single machine.

    Problem: The author was overwhelmed by walls of text from parallel Claude Code sessions, which limited how many sessions could be steered in parallel to about five.

    Posted by @sadilek · Original comment ↗

    “I'm building an open-source "spatial workspace for directing AI coding agents like an engineering director".”
  2. Obeya

    Open-source spatial workspace for directing AI coding agents like an engineering director. It provides a canvas per project with work laid out as cards, one agent running in a git worktree per card, and chat condensed to WhatsApp-style back-and-forth instead of a litany of tool calls. Agents typically produce a narrated demo video that the user can comment on by voice to have the agent go another round, with remote control from a phone and remote workers to scale beyond a single machine.

    Problem: Walls of text from parallel Claude Code sessions overwhelmed the author, limiting how many sessions could be steered in parallel to about 5 or 6.

    Posted by @sadilek · Original comment ↗

    “I'm building an open-source "spatial workspace for directing AI coding agents like an engineering director". I got overwhelmed by walls of text of parallel Claude Code sessions. This limited the sessions I could steer in parallel to maybe 5 or 6. I knew there had to be a better way, and I found one: A canvas per project. Work laid out as cards. An agent in a worktree per card. Chat condensed to WhatsApp-style back and forth, not a litany of tool calls. Result is usually a narrated demo video that I can comment on by voice to have the agent go another round. Remote-control from my phone. Remote workers to scale beyond my machine.”