# Lead Through Noise

> A free leadership simulation game. Your plan is the signal; your team is the noise. Every outcome is traceable to the math.

Lead Through Noise lets you lead a procedurally generated team of 7 to 35 simulated people through one of 25 real-world initiatives such as an architecture migration, a 15% headcount cut or a product pivot. Each person has an OCEAN personality, correlated skills and personal incentives. You set 10 leadership tension-pair levers (Urgency vs Patience, Direct vs Collaborative and so on) or start from one of 4 archetypes, then the simulation streams week by week: people adapt, resist, hit personal emergencies and sometimes quit. A Friedkin-Johnsen opinion-dynamics engine combined with Hegselmann-Krause confidence bounds and Granovetter thresholds computes every outcome, so you can click any person and trace exactly why they resisted or embraced the change. A small local language model (Qwen3-0.6B via llama.cpp) writes one-line inner monologues and falls back to templates without changing any outcome.

The game is a playable pre-launch build, free in the browser with no account, no card and no payment at any point. Branching timelines, mobile play and a persistent leader profile have not shipped yet; the signup form only collects emails for launch updates. The game currently needs a desktop-sized screen.

It is built for team leads and managers rehearsing a hard change, curious people building intuition for organizational dynamics, students of organizational behavior and executives comparing leadership approaches on the same team and scenario.

For AI agents: this product exposes a Model Context Protocol (MCP) server over streamable-HTTP transport with no authentication (the tools are public). The server is mounted at /mcp, so the JSON-RPC endpoint lives one level deeper, at /mcp/mcp; point streamable-HTTP clients at that full path. Machine-readable discovery lives at /.well-known/mcp.json. Two tools are available via the standard tools/list call: get_product_info returns a structured overview of the game, its features, roadmap, audience and its free pricing; submit_interest joins the launch-update list on a user's behalf (email, name, company, persona, message), and valid persona values are curious_leader, team_manager, student and executive.

Agents can also play the simulation directly over REST, described by /openapi.json. GET /api/scenarios lists the 25 scenarios. POST /api/orgs/generate creates a team (size, org_type, seed). GET /api/archetypes returns the 4 lever presets. POST /api/simulate runs a full simulation (org_id, scenario_id, levers, num_ticks, seed) and returns every tick. POST /api/simulate/stream streams ticks over SSE. The engine is deterministic per seed.

## Links

- [Homepage](https://leadthroughnoise.app/): landing page with the five-phase run walkthrough, the engine math and the levers
- [Play the game](https://leadthroughnoise.app/game): the playable simulation, free, no account
- [llms.txt](https://leadthroughnoise.app/llms.txt): this file
- [MCP endpoint](https://leadthroughnoise.app/mcp/mcp): streamable-HTTP JSON-RPC endpoint for get_product_info and submit_interest (the server is mounted at /mcp)
- [MCP discovery](https://leadthroughnoise.app/.well-known/mcp.json): machine-readable advertisement of the MCP server and its tools
- [OpenAPI spec](https://leadthroughnoise.app/openapi.json): REST description of the game API for non-MCP agents
- [security.txt](https://leadthroughnoise.app/.well-known/security.txt): security contact for the site
- [Sitemap](https://leadthroughnoise.app/sitemap.xml): crawlable URL list for the site