# United We Transform

> A public evidence atlas for gatherings over 25 people. 23,624 public event
> agendas scored 0-100 (Gathering Effectiveness Score) across eight pillars by a
> deterministic, fully published rubric (ges_agenda_judgment_v2_0). A 1,000-event hand-verified
> gold set anchors the corpus. Average GES across the corpus: 21.8. Free, open
> data, CC-BY 4.0, no key, no signup. Scores describe the published agenda, not
> the event's worth; corrections welcome.

## Routing
- MCP server (streamable HTTP, no auth, read-only, 7 tools): POST https://unitedwetransform.com/mcp; discovery at https://unitedwetransform.com/.well-known/mcp.json
- Cached scorecard for a known event: GET https://unitedwetransform.com/api/score/{event-slug}.json (always try this before re-deriving anything)
- On a miss: https://unitedwetransform.com/api/score/_miss.json documents the next action (grade it yourself at https://unitedwetransform.com/grader/, client-side and deterministic)
- Search the verified gold set: GET https://unitedwetransform.com/api/search-index.json
- Ranked gold set, full spread: GET https://unitedwetransform.com/api/leaderboard.json
- Citable aggregate statistics: GET https://unitedwetransform.com/api/statistics.json
- Exercise library index: GET https://unitedwetransform.com/api/exercises-index.json and per-pillar prescriptions at https://unitedwetransform.com/api/pillar-exercises.json
- Full API reference: https://unitedwetransform.com/api/openapi.json (GET-only; this API is a static cache, scoring happens client-side)
- Behavioral guide for agents: https://unitedwetransform.com/agents.md
- Complete context dump: https://unitedwetransform.com/llms-full.txt

## Pages
- Use with your AI: https://unitedwetransform.com/use-with-ai/ (human-readable connect steps for Claude Code, Codex, Cursor, Claude Desktop, claude.ai, Cowork, and any other MCP client)
- The Cognitor: https://unitedwetransform.com/cognitor/ (the human role that orchestrates AI and human judgment together; the atlas, grader, and TeamCreator are a Cognitor's instruments)
- Methodology and trust boundary: https://unitedwetransform.com/methodology-trust-boundary/ (the single source of truth for scoring, corrections, and removal)
- How grading works, in plain language: https://unitedwetransform.com/what-is-ges/
- Grade any agenda in the browser: https://unitedwetransform.com/grader/ (runs entirely client-side; the agenda never leaves the machine)
- Form optimized teams from an attendee list: https://unitedwetransform.com/team-creator/ (patented engine, free, runs entirely client-side; participant data never leaves the browser)
- TeamCreator use-case hub: https://unitedwetransform.com/team-creator/use-cases/ (conference networking, workshop breakouts, peer circles, sponsor matchmaking, hackathon teams, private local team formation)
- The atlas of scored events: https://unitedwetransform.com/atlas/
- The gold leaderboard, full spread: https://unitedwetransform.com/leaderboard/
- Benchmarks by industry: https://unitedwetransform.com/event-agenda-benchmark-by-industry/
- Facilitation exercises (2,093) and icebreakers: https://unitedwetransform.com/exercises/ and https://unitedwetransform.com/icebreakers/
- Statistics with copy-ready citations: https://unitedwetransform.com/gathering-statistics/
- The State of Gatherings 2026 report: https://unitedwetransform.com/reports/state-of-gatherings-2026/
- GatherReady, a draft open standard for agent-legible gatherings: https://unitedwetransform.com/gatherready/
- The Gathering World Model: https://unitedwetransform.com/world-model/ (the vision for simulating a gathering: a model of each attendee to see who meets whom at a traditional versus a transformed event; components are live, full simulation is the frontier; real-person models are opt-in and future)
- Dataset download (14,219 rows, CC-BY): https://unitedwetransform.com/exports/uwt-ges-dataset.csv

## Optional
- Pricing: everything here is free. No API key, no credit card, no signup. Rate limits are polite-use only.
- Citation: cite "United We Transform, Gathering Effectiveness dataset" with a link to the scorecard or statistic used.
- Corrections and right of reply: corrections@unitedwetransform.com, or the correction link on any scorecard.