https://github.com/machinepulse-ai/world2agent
World2Agent(W2A) is an open protocol that standardizes how Al agents perceive the real world.
https://github.com/machinepulse-ai/world2agent
agent proactive-agent protocol
Last synced: about 1 month ago
JSON representation
World2Agent(W2A) is an open protocol that standardizes how Al agents perceive the real world.
- Host: GitHub
- URL: https://github.com/machinepulse-ai/world2agent
- Owner: machinepulse-ai
- License: apache-2.0
- Created: 2026-04-23T08:49:29.000Z (about 2 months ago)
- Default Branch: main
- Last Pushed: 2026-04-27T13:34:58.000Z (about 2 months ago)
- Last Synced: 2026-04-27T15:07:53.931Z (about 2 months ago)
- Topics: agent, proactive-agent, protocol
- Language: TypeScript
- Homepage: https://world2agent.ai
- Size: 4.71 MB
- Stars: 39
- Watchers: 1
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- Contributing: docs/CONTRIBUTING.md
- License: LICENSE
Awesome Lists containing this project
README
Agents can't act on what they can't perceive.
Website ·
Quick Start ·
Sensors ·
SensorHub ·
Docs ·
Community
***
## What is World2Agent?
World2Agent (W2A) is an open protocol that standardizes how AI agents perceive the real world. Install a sensor, your agent gets structured, real-time data. Swap sensors freely — they all speak the same schema.
W2A isn't a product. It's an open protocol and an invitation. We built the first sensors — the real breakthroughs will come from the community.
→ [Why W2A? Full story](./docs/why-w2a.md)
## Architecture
**World → Sensor → Agent**
Sensors watch data sources and emit structured data following W2A Protocol. Your agent receives signals and decides what to do.
→ [Signal format spec](./docs/signal-format.md) · [Architecture deep dive](./docs/architecture.md)
## Quick Start
The fastest way to feel W2A is with Claude Code. In an active session, install the `world2agent` plugin:
```
/plugin marketplace add machinepulse-ai/world2agent-plugins
/plugin install world2agent@world2agent-plugins
/reload-plugins
```
Add a sensor — for example, Hacker News stories, frontier AI lab posts:
```
/world2agent:sensor-add @world2agent/sensor-hackernews
/world2agent:sensor-add @quill-io/sensor-frontier-ai-news
```
→ Browse the full catalog on [SensorHub](https://world2agent.ai/hub).
Restart Claude Code with the plugin channel loaded so sensor signals flow into your session:
```bash
claude --dangerously-load-development-channels plugin:world2agent@world2agent-plugins
```
> **Security — install only sensors you trust.** A sensor's signals drive what your agent perceives and does, so an untrusted sensor is effectively an untrusted instruction source. Stick to open-source sensors from authors you trust, and review the code first.
Native integrations for other agent runtimes are *coming soon*.
**Integrating W2A into your own agent system?** See the [developer quick start](./docs/quick-start.md#option-2-code--sdk--sensor) for the SDK code path.
→ [Full guide](./docs/quick-start.md) · [Multi-sensor](./docs/multi-sensor.md) · [SensorHub](./docs/sensorhub.md)
## Sensors
### [SensorHub](https://world2agent.ai/hub)
**SensorHub is the catalog of every W2A sensor** — official and community-built, organized by what each one perceives (markets, news, production alerts, weather, AI labs, …). Browse it to see what's available, view each sensor's signal schema, and grab the one-line install command. **Looking for a sensor? Start here.**
Every sensor is a standard npm package — `npm search w2a-sensor` works as a fallback if you prefer the raw view.
→ [SensorHub guide](./docs/sensorhub.md)
### Missing a sensor?
[Build your own](./docs/build-a-sensor.md) in ~50 lines. The `build-w2a-sensor` skill walks an AI coding agent through discovery, signal design, scaffolding, and the install recipe — install it with:
```bash
npx skills add https://github.com/machinepulse-ai/world2agent/skills/build-w2a-sensor
```
Once it's ready, ship it to npm:
```bash
npm publish
```
That's it — your sensor is now installable by anyone, anywhere.
## Roadmap
* **Graph layer** — compose and enrich signals from multiple sensors before they reach your agent. → [RFC](./docs/rfc-graph.md)
## Contributing
* 🔧 **Build a sensor** — `npm publish` and it's live
* 🐛 **Report bugs** — [open an issue](https://github.com/machinepulse-ai/world2agent/issues)
* 💡 **Suggest a sensor** — [Discussions](https://github.com/machinepulse-ai/world2agent/discussions)
→ [Contributing guide](./docs/CONTRIBUTING.md)
## Community
[Website](https://machinepulse.ai/) · [X / Twitter](https://x.com/MachinePulse_AI) · [YouTube](https://youtube.com/@MachinePulse_AI) · [Discord](https://discord.gg/hDjaD8pX)
## License
[Apache 2.0](./LICENSE)
***
Built by MachinePulse · Open source, open protocol, open invitation.