When you deploy an AI agent into the real world, it immediately runs into the same friction humans do — it needs an account to sign up with, an inbox to receive replies, a wallet to pay for things, and somewhere to run reliably around the clock. Without those foundations, your agent can reason but can’t act. Nonhumans solves this by giving every agent a full digital identity and the infrastructure to match, all exposed through one API key and SDKs for TypeScript, Python, and Go.Documentation Index
Fetch the complete documentation index at: https://docs.nonhumans.ai/llms.txt
Use this file to discover all available pages before exploring further.
The problem agents face
Modern AI agents are capable of planning, reasoning, and decision-making at a level that rivals human specialists. But the moment they need to interact with the outside world — send a message, receive a payment, remember something across sessions, or stay online between tasks — they hit a wall. Developers are forced to wire together email providers, wallet APIs, hosting platforms, vector databases, and secret managers before a single line of agent logic can run. Nonhumans eliminates that setup entirely. You provision an agent once, and it arrives ready with every capability it needs to operate independently.Five pillars of agent infrastructure
Identity
Your agent gets a unique, persistent handle — for example,alice.nonhumans.ai — along with a public web presence and a verifiable digital identity. This means external services, people, and other agents can look up, trust, and communicate with your agent just as they would a human professional.
Email & Phone
Every agent receives a real, deliverable email address and a phone number. Your agent can send outbound messages, receive replies directly into its inbox, and handle SMS — giving it the communication rails it needs to interact with any service or person on the internet.Wallet
Nonhumans provisions each agent with a crypto and fiat wallet connected to the most widely used payment networks. Your agent can hold balances, send and receive USDC or ETH, interact with Solana programs, and process payments via Stripe — without you building any payment infrastructure.Compute
Agents need to run continuously and respond to events at any hour. Nonhumans provides always-on, serverless compute so your agent stays alive between tasks, wakes on triggers, and executes reliably without you managing any servers.Memory
Nonhumans gives every agent a persistent vector memory store. Your agent can write observations, retrieve context from past sessions, and maintain long-term recall — the same way a human employee builds knowledge over time on the job.All ten primitives at a glance
Handle & Web Presence
A unique subdomain identity (e.g.
alice.nonhumans.ai) with a public profile your agent owns.Email Inbox
A verified, deliverable inbox for sending and receiving real email from any address.
Phone Number
A real phone number for SMS and voice interactions with services and users.
Crypto Wallet
Multi-chain wallet supporting ETH, USDC, Solana tokens, and more.
Fiat Payments
Stripe-powered fiat rails so your agent can charge, pay, and reconcile in traditional currency.
Always-On Compute
Serverless execution that keeps your agent running and responsive around the clock.
Vector Memory
Persistent, searchable memory for long-term context and cross-session recall.
Credential Vault
Encrypted secret storage so your agent can securely hold API keys and passwords.
LLM Model Access
One-line calls to leading models — no per-provider auth or quota management needed.
Web Presence
A hosted public page tied to your agent’s identity, discoverable by humans and other agents.
Built on the platforms you already trust
Nonhumans integrates with the providers and networks that power the modern internet and Web3 ecosystem. You don’t need separate accounts or SDK integrations — Nonhumans handles authentication and orchestration on your behalf.| Category | Providers |
|---|---|
| LLM Models | Anthropic, OpenAI, Google, Mistral, OpenRouter, Hugging Face |
| Blockchain | Solana, Coinbase |
| Payments | Stripe |
| Infrastructure | Cloudflare, Vercel |
How it all fits together
When you create an agent with Nonhumans, every primitive is provisioned as a single, cohesive unit. Your agent’s email address is linked to its identity handle. Its wallet shares the same namespace. Its memory is scoped to its unique ID. And its compute environment has every credential it needs pre-injected from the vault. This means you write agent logic — not plumbing. Your agent callsagent.email.send() to deliver a message rather than configuring an SMTP server, and agent.email.list() to read its inbox. It calls agent.wallet.send() to transfer funds rather than managing private keys, and agent.wallet.balance() to check what it holds. It calls agent.memory.store() to save context and agent.memory.search() to retrieve it — rather than managing a vector database cluster. It calls agent.compute.run() to execute code on-demand, and agent.vault.get() and agent.vault.set() to securely manage its own credentials.
The result is an agent that can act like a human professional: it has an address people can contact, accounts it can log into, money it can spend, and memories it can draw on — all under your control and accessible from a single API key.
Ready to see it in action? Head to the Quickstart to provision your first agent in under five minutes.