EHExplain Him
Menu

Explain Him

Express your idea once. Explain Him explains it to everyone.

Give people one page. Their own AI agent explains it at the depth they need, in the conversation they already use.

sectionHow it works

One shared explanation

How it works

The author publishes the meaning once. Each person can then explore it through the AI agent they already know.

One page

Give the link to your AI agent.

Ask in your own words. The agent understands your question, answers at the right depth, and uses this page as the visual surface.

1familiar conversation

Your agent is the conversation. Explain Him is the page.

There is no second chat to learn. The answer stays in the familiar agent interface, while the page gives that answer a shared visual place.

The complete path

One idea, six moments

Author → person → AI agent → personal view

  1. 01AuthorAuthor

    Publishes one page with the idea's canonical meaning and the sources that support it.

  2. 02PersonPerson

    Gives the page link to the AI agent they already use and asks a question.

  3. 03AIAI agent

    Understands the person's question and how much depth will be useful.

  4. 04AIAI agent

    Answers in the familiar agent conversation. There is no second chat.

  5. 05AIAI agent + page

    When a visual would help, adapts the page locally for that person.

  6. 06EHExplain Him

    Keeps the authored original unchanged. Personal explanations are saved separately in the browser.

The agent stays the conversation. Explain Him stays the visual surface.
Missing detail→Question to author

If the idea does not say, the agent asks.

It never invents missing information. With the person's approval, the question goes back to the author.

For authors and readers

How to express your idea

Publish the meaning and its evidence clearly. The agent can choose the best explanation path without changing what you meant.

Readers can trust where an answer came from.

Material claims stay connected to the authored page and its sources. The agent tells the reader whether something is current, planned, or still an open question.

Meaning belongs to the author

The published page remains the canonical explanation.

Sources stay visible

Readers can trace important claims back to their origin.

Gaps become questions

Missing information goes back to the author, never into a guess.

Personal means yours, not rewritten.

The original stays unchanged. Your explanations are saved separately in this browser, so another person can follow a different path without altering the author's page.

For developers: how it works internally

These diagnostics describe the page runtime. They do not mean an AI agent is connected or has called the page tools.

Authored HTMLCanonical page
+
Operation logIndexedDB
=
Visible DOMSafe typed blocks

WebMCP runtime status

Protocol v5 exposes one direct page tool: explain_tool. A verified page status confirms registration, not an agent connection.

Page WebMCP API – checking

Runtime boundaries

  • Typed schema: callout, comparison, workflow, timeline, diagram.
  • Browser fallback uses the same local workspace API.
  • Provenance keeps source path, status, and optional ref.

Test harness

Runtime test controls

0 saved test blocks

Personalized · 0 saved local presentations

Local history