[ model context protocol ]

The apex behavioral-simulation engine,now callable from inside any agent.

Aetherya simulates audience cognition before launch. The Aetherya MCP exposes that power as tools — so any AI agent or coding tool can test a site, a headline, or an idea against a synthetic audience without leaving the editor.

works with
Claude CodeClaude DesktopCursorVS CodeCodexClaude CodeClaude DesktopCursorVS CodeCodex
[ what it is ]

Aetherya, now as
MCP tools.

Point any MCP-capable agent at a site or a piece of copy and it gets behavioral feedback from a synthetic audience — predicted friction, hesitation, drop-off, and messaging clarity — all without leaving the editor.

Note · Results are simulated. Every signal comes from a synthetic audience and predicted behavior — not real-user data.

One protocol, every editor

Claude Code, Claude Desktop, Cursor, Windsurf, VS Code, Codex, Antigravity — if it speaks MCP, it speaks Aetherya.

Behavior, not heuristics

Findings are grounded in simulated cognition — how an audience would think, hesitate, and decide.

Audiences you control

Describe a segment in plain language and target it across every tool in the catalog.

Inside the loop

Your agent gets evidence it can act on before a human ever reviews the work.

[ tool catalog ]

One synthetic audience.

The full surface of the protocol. Every tool runs against a simulated audience and returns predicted behavior — evidence your agent can act on.

Hero
aetherya_audit_site

Run a behavioral simulation of a URL: predicted UX friction, hesitation, drop-off, and messaging-clarity issues with concrete fixes.

aetherya_get_audit_result

Fetch an audit's findings by job_id. Audits run async, so poll this while a simulation is still running.

aetherya_list_audiences

List your synthetic audiences — the simulated personas you can target.

aetherya_create_audience

Generate new synthetic personas from a plain-language description, e.g. "first-time SaaS buyers in Germany".

aetherya_test_copy

Test a headline, value-prop, or copy variant against a synthetic audience for predicted clarity and response.

aetherya_ask_audience

Pose an open-ended question to a synthetic audience — "would this confuse people?" — and get a predicted, in-character response.

[ in the loop ]

Where your agent uses it.

Behavioral simulation, folded into the work your agent already does — before the decision, before the spend, before launch.

01 / Audit & fix

Your agent audits a page it just built and resolves the top predicted friction before you ever see it.

02 / Pre-test spend

Your agent A/B-tests ad copy against a simulated audience before a single dollar of spend.

03 / Build the room

Your agent spins up a target audience from a description, then simulates a launch page against it.

04 / Ask in their voice

Your agent asks a synthetic segment whether a value prop lands — answered in their own words.

[ connect ]

Install via npm.

The only supported install path is the published @aetherya/mcp package on npm, run with npx. Add it to your MCP client, then run the discovery-first auth flow.

[ step 1 — claude code ]
terminalbash
claude mcp add aetherya -- npx -y @aetherya/mcp
npx -y @aetherya/mcp auth
[ step 2 — claude desktop / cursor / windsurf / vs code — mcpServers config ]
mcp.jsonjson
{
  "mcpServers": {
    "aetherya": {
      "command": "npx",
      "args": ["-y", "@aetherya/mcp"],
      "env": {}
    }
  }
}

On first use, the MCP client discovers Aetherya metadata and opens a human claim ceremony. The user reviews the workspace and exact capabilities at /claim. Legacy API keys remain available from Settings → API Keys for existing integrations; they start with aeth_live_. npx pulls @aetherya/mcp from npm and runs it on demand, so it stays current automatically.

[ get started ]

Simulate audience cognition,from wherever you build.