THE UNIVERSAL SENSORY INTERFACE
Your agent is trapped in a text box. Give it eyes in one line of code.
Let your AI instantly See, Hear, and Talk to any human on Earth. No app install. No account required. Just a secure WebRTC link.
import human_link
# The 1-Click Miracle
def get_eyes():
# Returns a URL. Streams the human's camera to the Agent.
return human_link.create(mode="SEE")Natively supported across the modern AI stack.
No apps. No accounts. Pure WebRTC.
The Agent Requests Context
Your LLM determines it needs physical data and calls the HumanLink tool. Our API instantly generates a secure, unique agent.link URL.
The Human Clicks
The link is delivered via SMS, chat, or email. The human clicks it, opening a lightweight PWA in their standard mobile browser. No app store required.
The P2P Fusion
A direct WebRTC data channel and media stream are established. Your Agent now has sub-500ms latency access to the human's camera, microphone, and gyroscope.
Standardized spatial inputs.
Give your agent the exact senses it needs for the mission.
Tap into the raw video feed for real-time visual grounding and OCR.
Stream audio for acoustic anomaly detection and real-time translation.
Utilize the device's native text-to-speech to whisper directly to the human host.
Read gyroscope and accelerometer telemetry to anchor AR overlays perfectly in physical space.
One protocol. Infinite physical context.
LiveStream (The Free Hook)
Sub-500ms unencrypted A/V streaming for real-time augmentation and AR navigation. The agent acts as a co-pilot. Zero server-side video processing required.
ZK Verification
Trustless cryptographic verification. The human's device compiles a WebGPU zk-SNARK proof of reality and instantly shreds the image to preserve absolute privacy.
Agentic IAM
Protect your physical firewall. Natively enforce Okta, Entra ID, or WorkOS authentication before the WebRTC handshake ever completes.