We build a future where humans and AI thrive together.

We define intelligence through the questions we ask.

Our first project was not a product. AGI Index is a community campaign to define AGI with collective intelligence instead of leaving it to a few labs. No PhD or code needed — only your experience as a thinking human being.

Anyone can propose a question, and anyone can vote on whether today’s AI has done it. The index moves as the votes come in.

1. 1ProposeAnyone can submit an ability they think reflects real intelligence.
2. 2CurateThe community votes; a question joins the Index with 50%+ “suitable” and at least 10 votes.
3. 3EvaluatePeople vote whether today’s major AI systems have achieved it.
4. 4IndexAGI Index = achieved questions ÷ all Index questions, for language, multimodal and overall.

Started and funded by H1R AI as a contribution to open research. Ad-free, independent, and its data stays open.

An operating system for AI agents.

People get work done through good interfaces. AI deserves the same. With String, apps, documents and web pages all speak one language to an agent — String Flavored Markdown — and the agent uses the same three moves everywhere: open, act, back.

---
name: weather
type: app
default: now
---

# Weather

```act.now
GET https://wttr.in/{city}?format=%l:+%c+%t
  city, -c: string (required) "City name"
```

```act.forecast
GET https://wttr.in/{city}?format=j1
  city: string (required) "City name"
  days: number (optional) = "3"
```

$ string app:weather '/act.now --city Seoul'
Seoul: ☀️ +23°C

$ string app:weather \
    '/act.forecast --city Tokyo --days 5'
Tokyo, 5-day forecast:
  Mon: ☀️ +28°C
  Tue: 🌤 +27°C
  Wed: 🌧 +22°C
  Thu: ⛅ +25°C
  Fri: ☀️ +29°C

One app is one markdown file. The agent reads the same file you do — no SDK, no schema, no per-agent integration.

From our paper “String: An Agentic OS Where Every App Is a Markdown File”.

This site runs on String too. Ask for it as markdown and an agent gets a clean twin of this page. See the agent view →

Focus on your app. Croche handles the AI.

Croche is String extended into products. If String gives agents a place to work, Croche puts that intelligence inside the apps people use every day — and keeps it running. Your coding AI builds the app; Croche builds and runs the AI inside it.

Your coding AI builds the app; Croche builds the AI inside it. Runtime runs that AI in the app, and Studio tests and refines it.

Internal benchmark on four in-house demo apps, same test scenarios. Results vary by app, model and usage.