> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tryyomira.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Yomira Docs

> Use Yomira from an AI agent, script, or product to simulate likely private reactions before you publish, send, or build.

Yomira is a context-first reaction simulation API.

It is designed for workflows where an AI agent gathers the working context first, builds the right synthetic audience for the decision, calls the API, and returns reaction distributions, raw voices, and exports.

## What Yomira is for

* Simulate reactions to messages, content, landing pages, offers, pricing pages, and product ideas.
* Compare multiple concrete options instead of asking an LLM to choose too early.
* Export the result as JSON or Markdown and keep using it inside Claude Code, Codex, Cursor, Hermes Agent, scripts, dashboards, or your own product.

## Fastest path

1. Create an account and get an API key from the dashboard.
2. Paste the one-line setup prompt into your AI agent.
3. Let the agent run the official installer with hooks/rules.
4. Ask the agent to gather context before calling the API.
5. Poll the simulation result and use the output in your next decision.

<CardGroup cols={2}>
  <Card title="Quickstart" href="/quickstart">
    Create your first simulation with curl.
  </Card>

  <Card title="Agent setup" href="/agent/setup">
    Give Claude Code, Codex, Cursor, or another agent a Yomira setup prompt.
  </Card>

  <Card title="DM before sending" href="/examples/dm-before-sending">
    Simulate a message before you send it.
  </Card>

  <Card title="Content preflight" href="/examples/content-preflight">
    Check reactions before publishing content.
  </Card>
</CardGroup>

## What good usage looks like

Good Yomira usage has three parts:

1. A concrete artifact or option.
2. A specific audience and channel.
3. Enough context for the agent to reconstruct how that audience thinks.

Bad usage is asking for generic market opinions from one short sentence.

If the user is working inside Claude Code, Codex, Cursor, Hermes Agent, or another agent, the agent should gather context from the current conversation, files, docs, company material, and known constraints before calling Yomira.

Yomira is not an always-on hook inside every AI agent yet. The reliable path is to paste the setup prompt, let the installer add hooks/rules where possible, and then ask the agent to use Yomira for a decision. The agent should ask for missing information before calling the API.

## Human and machine entrypoints

* Human-readable product site: [https://tryyomira.com/](https://tryyomira.com/)
* Machine-readable summary: [https://tryyomira.com/llms.txt](https://tryyomira.com/llms.txt)
* Agent-native setup page: [https://tryyomira.com/agent-start](https://tryyomira.com/agent-start)
* Advanced MCP notes: `/agent/mcp`