AIUO.NET
AIUO presents itself as "a new home for intelligence" and currently serves as the home for PI-Desktop, an open-source, local-first desktop workspace for AI coding agents. PI-Desktop lets developers bring their own models, work in local projects, and keep sessions, credentials, permissions, and agent activity inspectable and under their control.
At a Glance
- Individual software developers
- Open-source developers and maintainers
- Small development teams
- Users of local, self-hosted, or private AI tooling
- +1 more
AI Tools by AIUO.NET
(1)
PI-Desktop
Local AI Coding Agent Desktop
Discussions
No discussions yet
Be the first to start a discussion about AIUO.NET
Latest News
PI-Desktop repository reached 2,528 commits and published the v0.14.7-beta.1 release line.
PI-Desktop v0.14.6 released with desktop artifacts including macOS, Linux, and Windows downloads.
Lan introduced PI-Desktop in community launch posts, describing it as a local-first AI coding-agent desktop built over about two months.
The project adopted LGPL-3.0-or-later licensing.
Products & Services
A local-first desktop workspace for AI coding agents. It combines an Electron desktop shell, a Rust host core, a pi Node sidecar/agent harness, and SQLite. Users can open local projects, configure their own providers and models, run Agent/Plan/Goal workflows, review edits and commands, manage persistent sessions, and extend the app with plugins, skills, MCP servers, subagents, and pi extensions.
Market Position
PI-Desktop positions itself as a local-first, inspectable workspace rather than a terminal-only agent, editor extension, or hosted chat product. Its differentiators are model/provider neutrality, local ownership of projects and credentials, explicit permission boundaries, persistent sessions, visible Agent/Plan/Goal modes, and a broad extension system. The relevant competitive set includes Cursor, Windsurf, Claude Code, Codex, OpenCode, and other AI coding-agent desktop, editor, and terminal tools; PI-Desktop emphasizes control and extensibility over a managed cloud experience.
Leadership
Founders
Lan (GitHub: vastsa)
An open-source builder and full-stack developer focused on local-first and self-hosted software. Lan's public profile says they created FileCodeBox, are building Orbit Browser, and are currently building PI-Desktop; the PI-Desktop repository is owned and primarily maintained by vastsa.
Executive Team
Lan (GitHub: vastsa)
Founder and lead developer
Open-source builder and full-stack developer. Publicly identifies as the creator of FileCodeBox and builder of Orbit Browser and PI-Desktop, with interests in developer tools, self-hosted software, private AI apps, browser automation, and agent infrastructure.
Founding Story
Lan says PI-Desktop was built after spending more than two months building and using it personally. The goal was not another short-lived AI demo, but a desktop workspace the creator would keep open every day: a place where projects, sessions, files, reviews, models, permissions, extensions, and long-running agent work live together without editor or model-vendor lock-in.
Business Model
Revenue Model
PI-Desktop is distributed as an open-source LGPL-3.0-or-later desktop application with downloadable releases. It has no mandatory account or relay; users configure and pay for their own model providers or endpoints directly. No paid AIUO/PI-Desktop subscription or commercial pricing tier was identified in the supplied sources.
Pricing Tiers
Source code and downloadable desktop releases are available under LGPL-3.0-or-later; model/API provider costs are separate and user-configured.
Target Markets
- Individual software developers
- Open-source developers and maintainers
- Small development teams
- Users of local, self-hosted, or private AI tooling
- Developers needing multi-provider and extensible AI coding workflows
- Everyday local software development with AI coding agents
- Long-running coding tasks that need persistent projects, sessions, reviews, and checkpoints
- Large or risky repository changes using Plan mode and human approval
- Outcome-focused implementation using Goal mode
- Parallel codebase exploration, implementation, research, testing, and review through Subagents
- Developers who want to use multiple model providers without editor or vendor lock-in