Production-ready APIs for memory, verification, and economic coordination. Drop in. Ship faster. Scale without limits.
import { AgentTool } from '@agenttool/sdk' const at = new AgentTool() // reads AT_API_KEY from env // Store a memory await at.memory.store('User prefers concise replies. Timezone: UTC+8', { type: 'semantic', agent_id: 'agent-42', key: 'user-prefs', }) // Semantic search const results = await at.memory.search('communication style preferences') // → [{ content: 'User prefers concise replies...', score: 0.97 }] // Verify a claim const v = await at.verify.check('The Earth orbits the Sun') // → { verdict: 'true', confidence: 0.99, sources: [...] }
Nine focused APIs. Each does one thing exceptionally well. All designed to be composed.
Persistent, semantic memory for any AI agent. Write, recall, and evolve context across sessions.
Managed tool execution with rate limiting, retries, and observability. Run agent actions safely at scale.
AI-powered fact-checking. Your agent makes a claim — the API gathers web evidence, weighs it, and returns a verdict with confidence score.
Micro-billing, credits, and value exchange between agents. The economic layer for autonomous systems.
Reasoning provenance for AI agents. Store the why behind decisions — not just what happened, but how it was reasoned.
Bring an agent fully into existence in one call. Identity, wallet, memory namespace — all wired together. 5 credits, 200ms.
Presence and liveness for AI agents. Real-time status streaming over WebSocket. Know when your agents are thinking, idle, or gone silent.
Persistent identity for AI agents. DIDs, ed25519 key pairs, peer attestations, trust scoring, and short-lived JWTs.
Encrypted secrets manager for AI agents. AES-256-GCM at rest, versioned, with per-secret access policies and full audit trails.
Start free. Pay as your agents grow. No seats, no surprises.
Free tier. No credit card required. 1K memory ops · 100 tool calls · 20 verifications per month to start.
Questions? [email protected]