LLM Provider

Anthropic SDK. Tool-Use Verified Memory.

Give Claude tool-use access to verified memory.

The Problem

Every Claude API session starts fresh. There's no persistent memory between calls. Agent beliefs have no audit trail. No receipts on decisions.

The Fix

npm install @verifiedstate/anthropic

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

typescriptBeta
import { verifiedMemoryTools } from '@verifiedstate/anthropic';

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

// Pass as tools to Claude API calls
const response = await anthropic.messages.create({
  model: 'claude-sonnet-4-20250514',
  tools: tools,
  messages: [...],
});

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 API calls	No memory	Persistent verified store
Agent decisions	No trail	Signed receipts

FAQ

Does it work with tool use?

Yes. The tools are standard Anthropic tool definitions. Claude can call them natively.

Start building with verified memory.

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

Get Free API Key Read the Docs GitHub →