AI Connect

AI

Live state for agents — analytics, training, and automation.

Connect application states and human-curated data directly to AI agents. The same foundation as Applications and Humans — one live source agents can read, subscribe to, and act on.

Create API KeyAI docs

Why BigState for AI

Live context for agents — without a separate export pipeline

⚡

Live Data Source

Agents always work from the latest versioned state — not a stale export or overnight dump.

🔗

Same API as Your Apps

Reuse the objects, policies, and deliveries your services and UIs already use. No parallel data path for AI.

📦

Metrics and Files

Feed structured metrics inline, or pull files and images when state points at richer payloads.

🛡️

Scoped Agent Access

Give each agent credentials limited to the streams it needs — nothing more.

How It Works

From live state to agent action in three steps

1

Connect

Authenticate the agent as its own principal with an API key or token.

const httpClient = new BigStateHttpClient({ apiKey });
2

Read

Fetch the latest state — or subscribe so the agent reacts as values change.

const state = await httpClient.stateGet(OBJECT_NAME);
3

Act

Use the data in your agent runtime — tool calls, workflows, analytics, or training jobs.

// Agent decides from live state

Wire an Agent in a Few Lines

Read latest state — or subscribe and react as it changes

example.js
import { BigStateHttpClient, BigStateWsDeliveryClient } from 'bigstate.client.javascript';

const httpClient = new BigStateHttpClient({
	baseUrl: 'https://api.bigstate.dev',
	apiKey: 'AGENT_API_KEY',
});

const wsClient = new BigStateWsDeliveryClient({
	baseUrl: 'https://ws.delivery.bigstate.dev',
	apiKey: 'AGENT_API_KEY',
});
example.js
const METRICS_OBJECT = 'metrics:live@YOUR_OWNER';

const latest = await httpClient.stateGet(METRICS_OBJECT);
console.log('Latest metrics:', latest.value);

// Agent uses live state for a decision or tool call
await agent.run({ context: latest.value });
example.js
import { NotificationType } from 'bigstate.client.javascript';

wsClient.setDeliveries(['agent-metrics']);
wsClient.setMessageHandlers([(message) => {
	if (message.type !== NotificationType.Message) return;
	agent.onStateUpdate(message.data.state);
}]);

Built for Agent Workloads

Training, analytics, automation — one live source

📈

Live Analytics

Point agents at live metrics and application state so insights stay current as the system changes.

🧠

Training Signals

Stream real-world state into training or fine-tuning pipelines instead of batch exports.

⚙️

Workflow Automation

Trigger tools and workflows when state updates — agents react to the same events your apps already publish.

🎯

Informed Decisions

Give agents curated human reports and live app data in one model — better context, fewer stale inputs.

Ready to Connect Agents?

Start from the State API or explore use cases — give agents the same live data your apps and people already share.