Build the audience first,
then simulate the reaction.

The problem

Most AI tools answer from the prompt in front of them. But real reactions come from people with context: what they already believe, what problem they have, how much they trust the sender, what alternatives they compare against, and whether the artifact reaches them at the right moment. A useful simulation has to reconstruct that audience before it predicts the reaction.

The product

Yomira is a context-first simulation API. In self-serve mode, the user's AI agent gathers the artifact, company context, channel, audience clues, constraints, and known worries, then Yomira builds a synthetic audience and runs reaction simulations. In enterprise mode, Yomira builds grounded audience datasets from real customer, market, and social data.

Self-serve vs enterprise

The line is data. Self-serve is for fast, context-enriched synthetic simulations from the context your AI agent can gather. Enterprise is for grounded simulations where Yomira helps build the audience dataset from real source material.

Use cases

• Content preflight — check how a post, article, ad, or GEO page may be received before it goes live.
• Message decisions — compare sales DMs, follow-ups, investor notes, user interviews, and sensitive outreach before sending.
• Product and offer positioning — test landing pages, pricing pages, feature explanations, and launch copy against likely buyer reactions.
• Venture idea checks — see which parts of a new business idea create desire, confusion, distrust, or concrete willingness to try.
• Embedded simulation products — add reaction simulation to your own workflow, research product, agency tooling, or decision system through the API.

Concrete workflows

How it works

1. 1. Collect working context

   Your AI agent extracts the artifact, company context, channel, audience clues, constraints, and known worries from the conversation, files, docs, or repo.

2. 2. Generate the audience

   Yomira turns that context into a synthetic audience that matches the decision instead of a generic persona list.

3. 3. Simulate individual reactions

   Each agent reacts to the exact artifact, producing private voices, likely actions, confusion, desire, and suspicion.

4. 4. Decide with evidence

   Yomira returns distributions, clusters, raw voices, missing-data warnings, and exports for the next decision.

Official skill

Install the Yomira skill into Claude Code, Codex, Cursor, Hermes Agent, or another AI agent. The skill tells your agent to gather the necessary context first, construct the target audience, and then call the real API instead of guessing from a thin prompt.

Install the official Yomira skill from this GitHub repo:
https://github.com/ryuzo-k/yomira/tree/main/skills/yomira

After installing, use it whenever I ask you to simulate reactions to a message, post, landing page, offer, pricing page, product idea, or decision option.

If I provide a Yomira API key, call the real API. Do not replace it with casual reaction prediction unless I explicitly ask you not to use the API.

What you get

"I get why this matters. I would try it if the setup is fast and the output shows real private objections, not just generic AI feedback."

Real example

We simulated a simple workflow: a user installs the skill, gives an API key to Claude Code or Codex, and asks it to check a sales DM before sending. The API ran 40 simulated people.

"This could shortcut endless A/B testing if accurate." Another voice said: "Wish this was a plug-and-play tool, not a skill install."

Simulation id: fe1d26c8-8a74-413c-b9bf-3b0b6ab7e289

What early users are teaching us

Who it is for

Start with builders, founders, marketers, agencies, and agent-native users who already make decisions inside AI tools. Self-serve is for context-enriched simulations from a user's workspace. Enterprise is for grounded simulations where Yomira constructs the audience dataset from customer, market, and social data.

Pricing

Self-serve credits are for context-enriched synthetic simulations. Enterprise work is custom when you need Yomira to build grounded audience datasets from CRM data, customer lists, reviews, interviews, X/social data, or private market context.

API quickstart

curl -s -X POST "https://tryyomira.com/api/simulate" \
  -H "content-type: application/json" \
  -H "x-api-key: $YOMIRA_API_KEY" \
  -d '{
    "objective": "Should we publish this landing page?",
    "artifact": { "type": "landing_page", "content": "..." },
    "audience": { "description": "founders, marketers, agencies, and AI-agent users" },
    "simulation": { "target_n": 100, "mode": "standard" }
  }'

Mora

Mora is a free companion skill. It maps the possible options before you simulate them. Use it when you do not yet know which messages, positioning paths, launch plans, or product directions are worth testing.

FAQ