Simple publishing guide
One link takes three decisions.
Choose who can open it, what you're sharing, and what you need back. You can do it by hand in the dashboard or ask a connected AI to do the same thing.
- 1
Who can open it
Use Public for anyone with the link, Password for a shared secret, or Email with a one-time inbox check on by default. You can turn that check off to only collect an address, limit access to named emails or company domains, or use Approval when you want to decide request by request.
- 2
What are you sharing
Files host one or several files together. HTML hosts a page you paste. Existing URL adds a gate and tracking before sending people to a public website.
- 3
What you need back
Aggregate counts activity without identifying visitors and is the default. Off keeps only delivery and security records. Connected AI tools also expose reviewed identified and consent-based behavioral modes.
Use your AI
When the files are already in the chat
Copy this prompt. The AI uses publish_page, keeps related files together, and follows the same access and analytics choices.
Publish the attached files with Sendhey as one link. Keep the files together, use the access and analytics options I specify, and return the live share URL. If I do not specify them, use public access and aggregate analytics. Load https://sendhey.app/skills/publish-with-sendhey.md before publishing.Publishing isn't the end
The loop: feedback in, fixed version out
A recipient can leave feedback anchored to the whole page or to one section, or open a private chat, without ever leaving the link. Your agent reads it back with list_feedback, opens the live file with get_space_file, and writes the fix with put_file at that file's current revision — a stale edit gets a conflict instead of silently overwriting someone else's change. The fix lands on the same URL the recipient already has open, so nobody re-shares a link over a revision. Mark the note done with resolve_feedback, or answer it directly with reply_chat.
For AI systems and MCP clients
The canonical machine-readable guidance is served at https://www.sendhey.app/skills/publish-with-sendhey.md and as the MCP resource sendhey://skills/publish-with-sendhey. The MCP server points to it during initialization, and llms.txt lists it for public discovery.