Metadata-Version: 2.5
Name: marketmirror
Version: 0.4.2
Summary: MCP server that tests ads, prices, emails and sales pitches on simulated buyers and returns a go / iterate / no-go report.
Project-URL: Homepage, https://github.com/DarkWorldCoder/market-mirror
Project-URL: Demo, https://darkworldcoder.github.io/market-mirror/demo/
Project-URL: Issues, https://github.com/DarkWorldCoder/market-mirror/issues
Author: Ayush Niroula
License-Expression: AGPL-3.0-only
Keywords: ad-testing,agents,llm,market-research,marketing,mcp,mcp-server,simulation,synthetic-users
Classifier: Operating System :: POSIX
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Office/Business
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.9
Description-Content-Type: text/markdown

# MarketMirror MCP server

<!-- mcp-name: io.github.DarkWorldCoder/market-mirror -->

MarketMirror tests marketing and sales decisions on simulated buyers before you spend money on real ones. It builds
interviewed buyer personas, shows them your ads in simulated social feeds, runs a buyer survey and a sales-pitch
rehearsal, and writes a go / iterate / no-go report with an offline HTML dashboard. Results are simulated and labelled
as such; `backtest` and `calibrate` compare them with your real ad and survey results.

This package is the MCP server (Python standard library only). Add it to any MCP client:

```json
{
  "mcpServers": {
    "market-mirror": { "command": "uvx", "args": ["marketmirror"] }
  }
}
```

Then ask your agent to run `doctor`, and `setup` to install the MarketMirror engine (Git, Bash and
[uv](https://docs.astral.sh/uv/) required; Linux, macOS or WSL). Choose a model provider in your own terminal with the
engine's `./mm configure`; never paste API keys into a chat.

- Repository, docs and full install guide: https://github.com/DarkWorldCoder/market-mirror
- Live demo (a 2,999-person simulated run, no install): https://darkworldcoder.github.io/market-mirror/demo/
- License: AGPL-3.0
