Skip to main content

About Sundial

A workspace for persistent AI agents.

What is Sundial?

Sundial is a workspace for persistent AI agents built by Long Horizon Research, Inc. Users create agents that can perform tasks, manage schedules, surface information, and communicate on their behalf. Each agent can reach users via SMS, sending personalized updates, daily briefings, task reminders, and conversational responses directly to the user's phone.

Agents work inside your files and tools — not next to them. They edit your actual documents with tracked changes, search and save results to your workspace, and show what they did. They run persistently across sessions, so your work keeps moving even when you step away.

Skills

Agents are powered by skills — portable SKILL.md files that give any agent new knowledge and workflows. The skill registry has thousands of community-built skills you can browse, install, and publish. Skills work with Claude Code, Cursor, Gemini, GitHub Copilot, ChatGPT, and any agent that reads markdown.

What you can do

  • Create persistent agents that work in real-time collaborative workspaces.
  • Receive SMS updates, briefings, and conversational replies from your agents directly on your phone.
  • Browse skills — search thousands of community-built skills by category or keyword.
  • Publish your own skills from the dashboard or push directly with the CLI.

Resources

  • Documentation — guides, CLI reference, and the SKILL.md spec.
  • agentskills.io — the open Agent Skills specification.
  • GitHub — open source projects and community contributions.

Get in touch

Email us at team@longhorizonresearch.com or find us on Twitter at @sundialhub.