Daniel Sadilek
Daniel Sadilek is an individual freelance full-stack builder and architect, not a company. His Obeya project is an MIT-licensed, open-source local workspace for directing AI coding agents as an engineering director: tasks appear as cards on a canvas, agents work in the background, and return narrated demos for human approval.
At a Glance
- Independent developers
- Small software teams
- Engineering leaders directing AI-assisted development
- Teams using GitHub, GitHub Enterprise, Forgejo or Gitea
- +1 more
AI Tools by Daniel Sadilek
(1)Obeya
Canvas Workspace for Coding Agents
Discussions
No discussions yet
Be the first to start a discussion about Daniel Sadilek
Latest News
Obeya demos documented a fallback model when Claude Code reaches its usage limit, plus Codex takeover and related agent/model controls.
Obeya shipped a fix making Ctrl+mouse-wheel zoom advance in bounded steps on Linux and Windows.
Obeya improved the canvas color-ring interaction so clicking a group selects it without the old broad aiming behavior.
The phone view was updated so Running shows only work an agent is actively doing; queued and snoozed cards move to On hold.
Products & Services
Open-source desktop/local workspace for coordinating AI coding agents. It provides a spatial canvas of task and project cards, per-card agent workspaces, dependency-aware scheduling, narrated demos or HTML artifacts for review, voice interaction, approval and landing to main or pull requests, and a phone view for monitoring and approval.
Static/public pages where approved feature demos are published with chapters and descriptions; Obeya can publish to a repository demo site or a personal GitHub Pages/Cloudflare site.
Market Position
Obeya differentiates from terminal-centric and text-only coding-agent workflows by making work spatial and persistent, coordinating multiple agents by dependency, and turning each handover into a narrated demo or visual artifact for approval. It is local and open source rather than a hosted coordination service, and is explicitly aimed at people who want to direct agents without routinely reviewing their generated code.
Leadership
Founders
Daniel Sadilek
Independent full-stack builder and architect working across hardware, cloud, web, mobile and AI. LinkedIn identifies him as a former Google Senior Staff Software Engineer, where he was tech lead and engineering manager for Google Calendar Backend and API (2009–2021); previously a PhD researcher at Humboldt University of Berlin working on model-driven technologies, language engineering and wireless sensor networks. He was also a founding engineer at Relay.app (2021–2024).
Executive Team
Daniel Sadilek
Independent developer, architect and creator of Obeya
Freelance full-stack builder and architect; former Google Senior Staff Software Engineer and Google Calendar backend/API tech lead and engineering manager; PhD from Humboldt University of Berlin.
Founding Story
Sadilek said he created Obeya after becoming frustrated with parallel coding sessions, walls of terminal text, and having to act as QA for coding agents. His initial vision was to be the agents' engineering director: give tasks by voice, let agents work in parallel, watch demo videos or presentation documents, provide feedback, and approve changes to land them on main.
Business Model
Revenue Model
Obeya is distributed as free, open-source, local software under the MIT License. The public materials describe it as running on the user's machine rather than as a hosted SaaS product; users supply or log into their own agent services such as Claude Code or Codex.
Pricing Tiers
Local software available from source or downloadable macOS, Windows and Linux builds; the Hacker News launch describes it as free, open-source and local.
Target Markets
- Independent developers
- Small software teams
- Engineering leaders directing AI-assisted development
- Teams using GitHub, GitHub Enterprise, Forgejo or Gitea
- Projects that need local/private execution and phone-based approval
- Solo developers and engineering directors coordinating multiple AI coding agents
- Building features without repeatedly reading generated code or switching among terminal sessions
- Voice-driven tasking, feedback and approval based on demos
- Parallel projects and workstreams with dependencies or potentially colliding changes
- Reviewing external GitHub pull requests with an agent before execution
- Remote agent execution over SSH and monitoring/approval from a phone