Agent Framework

OpenAI Agents + Verified Memory.

Tool-based verified memory for OpenAI agents.

The Problem

OpenAI's Agents SDK provides tools and handoffs. Memory is left to you. Each agent invocation starts fresh unless you wire up persistence yourself — and that persistence has no verification, no receipts, no conflict detection.

The Fix

npm install @verifiedstate/openai-agents

Beta: tagging-level integration. Ingestion works via @verifiedstate/sdk with OpenAI Agents-specific source tagging. Native OpenAI Agents features coming.

typescriptBeta
import { verifiedMemoryTools } from '@verifiedstate/openai-agents';

const tools = verifiedMemoryTools({
  apiKey: 'vs_live_...',
  namespaceId: 'agent_memory',
});

// Add to your OpenAI agent as tools
const agent = new Agent({
  name: 'my-agent',
  tools: [...myTools, ...tools],
});

What You Get Back

✓ VERIFIED

receipt.json
{
  "receipt_id": "rct_7f3a...",
  "assertion": "User prefers dark mode",
  "status": "verified",
  "confidence": 0.94,
  "signature": "ed25519:a8c3f1...",
  "verified_at": "2026-04-06T14:32:01Z",
  "chain_position": 42
}

Without VerifiedState vs. With VerifiedState

Feature	Without	With VerifiedState
Between invocations	No memory	Persistent verified memory
Agent handoffs	Context lost	Shared verified store
Audit	None	Signed receipts

FAQ

Does it work with function calling?

Yes. The tools are standard OpenAI function definitions. Works with both the Agents SDK and direct function calling.

Start building with verified memory.

Free: 50,000 assertions/month. No credit card. No trial expiration.

Get Free API Key Read the Docs GitHub →