The runtime for AI-run businesses.
Agent primitives, autonomous companies, and the infrastructure that lets AI agents do real work — incorporate, bank, ship, sell. Updates and deep-dives from the Naïve team.
Sole-proprietor phone numbers: provision SMS without an LLC
The /phone primitive now supports a sole-proprietor tier — provision a carrier-registered US number with passed KYC and an SMS OTP, no LLC formation or EIN required.
defineProject: strict, honest agent infrastructure as code
naive.config.ts gets a strict entrypoint. defineProject refuses any declared field nothing consumes, kits: becomes the canonical declare-now-instantiate-later block, and every team runs durable by default — hermes only when you say so.
Introducing /sandbox: Disposable micro-VMs for every agent
Give any agent an isolated Linux micro-VM in one call — run shell commands, read and write files, expose ports at public URLs, checkpoint the whole machine, and fork copies of it. Billed only for the CPU, memory, and disk it actually uses; sleeping and parked sandboxes are free. No cloud account, no Docker, no infra.
What Is AI Agent Infrastructure? Governed Profiles, Primitives, and Policy
AI agent infrastructure is the governed real-world layer agents need to act — per-tenant identity, primitives, cloud when needed, and policy at the tool boundary. How Naïve fits.
What Is an AI-Native Cloud? The Hosting Slice Agents Can Call
An AI-native cloud is the hosting and data-plane slice agents can provision by API — apps, Postgres, storage, functions, compute, queues — under the same governed identity as other primitives. Not a hyperscaler replacement.
Introducing /audio: give your agent ears and a voice
Transcription, speech synthesis, and native audio conversations across a managed catalog of speech models — one endpoint, automatic routing with fallbacks, exact-cost billing in Naïve credits, and no speech-provider accounts to manage.
Build a Gusto payroll and onboarding readiness agent
Multi-tenant Gusto readiness on Naïve: read-only payroll blockers and onboarding status, Account Kit tool allowlists, full connect lifecycle, client.llm summaries.
Introducing /payments: let your AI agent pay for x402 resources in USDC
Give each agent its own crypto wallet and let it settle x402 paywalls autonomously — quote the price, pay in USDC on Base, and reconcile every receipt, with the spend cap enforced inside the custody plane.
Introducing /wallet: a per-agent crypto wallet with a custody-plane spend cap
Give each agent its own onchain crypto wallet — read its address and balance, and, as an operator, fund it, transfer, set the spend policy, and sweep it. The private key never enters the agent runtime and the spend cap is enforced inside the custody plane.
Governed HR agents in your customers' existing Gusto companies
Connect each customer's Gusto company to a governed agent profile on Naïve — read-only payroll and onboarding readiness today, with kit-level tool policy and audit.
How Agents Run SEO Keyword Research Loops
Run a governed SEO keyword research loop per tenant with AgentSEO — volume, SERP competitors, and Labs insights via forUser, plan-metered and audit-logged.
How Agents Track Brand Mentions in AI Search
Track how ChatGPT, Claude, Gemini, and Perplexity mention your brand with AgentAEO — LLM responses, mention search, and AI keyword volume per tenant.
How to Build a Usage-Metered AI SEO SaaS
Build a multi-tenant AI SEO/AEO SaaS with Customer Billing: plans enforce quotas at execution time — 429 on overage, forbidden when a primitive is off-tier.
Audit Logging for AI Agent Actions
Audit logging for AI agent actions: one immutable trail per tenant user — who acted, what ran, approval outcome, and spend — from the gateway.
How Agents Manage Subscription Spend With Virtual Cards
Issue per-tool virtual cards for AI agents buying SaaS subscriptions — monthly caps on the card network, approval gates, and instant revocation on churn.
How Agents Receive 2FA Codes With a Phone Number
Give each tenant agent a dedicated US phone number for inbound SMS — 2FA codes scoped per customer and readable through the governed Naïve /phone API.
How Agents Use Virtual Cards for Event Bookings
Give each event agent a spend-capped virtual card for venue deposits, tickets, and vendor payments — hard limits and approval gates enforced at execution time.
How AI Employees Work: The Specialized Workforce Behind an Autonomous Company
AI employees are specialized agents scoped to one Company — own identity, inbox, and primitive whitelist. How a Naïve workforce is structured and dispatched.
How to Give an AI Agent a Virtual Card With a Spend Limit
Give an AI agent a virtual card with a spend limit: issue per tenant user, cap in the Account Kit, and enforce every charge at the gateway.
How to Set a Budget Cap on an AI Agent
How to set a budget cap on an AI agent: combine agent-profile budget with Account Kit card limits so spend is enforced server-side, not in the prompt.
Naïve Inside Your Agent Framework (Governed Tools, Any Orchestrator)
Naïve as the governed tool layer under any agent framework — keep your orchestrator; add identity, connections, caps, and approvals via agentTools() or MCP.
Naïve supports Browserbase for Agent Browser Automation
Naïve supports Browserbase for agent browser automation — cloud browser sessions via /browser on one governed identity, with signup, login, and vault-backed credentials.
Naïve supports Doola for Agent Incorporation
Naïve supports Doola for agent incorporation — US LLC formation through /formation on one governed identity, filed via Doola after KYC and checkout.
Naïve supports Resend for Agent Email
Naïve supports Resend for agent email — real inboxes and send/receive through /email on one governed identity, with domain verification and deliverability built in.
OAuth for AI Agents: Per-User Connections
OAuth for AI agents should be per-user: each tenant connects their own apps — scoped by Account Kit, stored as connections, not shared tokens.
Tenant Isolation in a Multi-Agent SaaS
Tenant isolation in a multi-agent SaaS: every resource attaches to one tenant user — verify forUser scoping and cross-tenant tests before production.
What Is an AI Agent Identity?
An AI agent identity is the governed record an agent acts as — tenant user, Account Kit, resources, and audit trail — not the model or the API key alone.
What Is an MCP Session?
An MCP session is a short-lived, per-user credential that scopes an agent runtime to one tenant's governed tools — not your workspace API key.
Why Consolidate Agent Infra on One Governed Identity
One governed identity per tenant beats a stitched vendor stack — cards, email, vault, connections, and KYC on the same account, enforced at execution time.
Delegated, Revocable Access with Per-User MCP Sessions
Per-user MCP sessions give AI agents delegated, revocable access — short-lived, kit-scoped credentials instead of your workspace key. Here's the pattern.
Hosted vs Bring-Your-Own Runtime for AI Agents
Hosted vs bring-your-own runtime for AI agents: Naïve is runtime-agnostic — governance applies at the tool-call boundary either way. Compare both paths.
How to Add Human Approval to an AI Agent
How to add human approval to an AI agent: gate sensitive actions on your Account Kit, surface pending approvals, and replay on approve. A step-by-step guide.
How to Revoke an AI Agent's Access Instantly
How to revoke an AI agent's access instantly: suspend the agent profile or revoke an MCP session so the governance gateway denies all further tool calls.
Spend Caps, Approvals & Revocation for AI Agents
Spend caps, human approvals, and instant revoke for AI agents must be enforced server-side — Naïve's governance gateway applies all three on every call.
The Governed Agent Profile: End-to-End Control for AI Agents
A governed agent profile is an AI agent with identity, spend limits, approvals, audit, and instant revoke — enforced on every action. Here's the full lifecycle.
Build an Agentic Collections Platform (Dunning That Runs Itself — Under Permission)
Build a multi-tenant collections agent on Naïve: cron runs dunning on schedule, memory tracks each debtor's stage, and Account Kits gate every action.
How to Build a Multi-Tenant AI Brand-Protection Agent
Build an agentic domain-defense platform that watches for lookalike and typosquat domains, investigates live impersonation with OpenRouter's web-search server tool, and registers defensive domains — where no registration ever spends money without passing an execution-time approval gate. Full build on the Naïve SDK: per-tenant identity, AccountKit capability tiers, and the /domains primitive.
Build an Agentic Mobile QA Platform: Isolated Cloud Phones, Approval-Gated Provisioning, and Metered Per-Tenant Billing
Build a multi-tenant AI mobile-QA platform on Naïve: each customer gets isolated cloud phones, agents run plain-English test flows on real devices, provisioning and runs are approval-gated, and every device-minute is metered per tenant — without ever holding a Mobilerun key.
Build an Agentic Data-Pipeline Platform (Where the Agent Can't Run Code Without Sign-Off)
Build a multi-tenant data-pipeline platform on Naïve where an LLM writes the transform and runs it in a per-tenant isolated container — but spinning up compute and opening a shell freeze for a human at execution time, and no agent ever holds a cloud key.
AI Agent Permissions, Enforced at Execution Time
AI agent permissions should be enforced server-side at execution time, not prompted. Here's how policy templates gate primitives, apps, and approvals.
Building AI Agents Into Your SaaS: The Multi-Tenant Playbook
Building multi-tenant AI agents into your SaaS means giving each customer an isolated, governed agent. Here's the full architecture and how to ship it.
How to Build an Agentic Appointment Reminder & No-Show Recovery Agent
Build a two-way SMS agent that texts appointment reminders, reads replies, and reschedules no-shows automatically — on a real US number your agent legally can't fake. Full build with Naive's phone, identity, and LLM primitives, with execution-time permission enforcement shown at every step.
How to Store API Keys for AI Agents Safely
How to store API keys for AI agents safely: keep the key out of the prompt, encrypt it per user, and let the agent use it server-side. A step-by-step guide.
Multi-Tenant AI Agents: One Agent Per Customer, Isolated by Default
Multi-tenant AI agents give every customer their own isolated agent. Here's the operator, tenant-user, and Account Kit model behind per-user isolation.
Secrets Your AI Agent Can Use but Never Sees
Secrets management for AI agents means letting an agent use a credential without exposing it to the model. Here's the per-tenant vaulting pattern that fixes it.
What Is an Account Kit?
An Account Kit is a reusable policy template that governs what a tenant user's AI agents can do — which primitives are enabled and which apps they can connect.
Build an Agentic Ad-Buying Platform (Where Every Agent Gets a Card With an Issuer-Enforced Ceiling)
Build a multi-tenant ad-buying platform on Naïve where each client's agent gets its own funded virtual card with a hard, issuer-enforced ceiling — the spend cap the prompt can't cross.
Build an Agentic Operator for Tools with No OAuth (Where the Agent Uses a Password It Can Never Read)
Build a multi-tenant agent on Naïve for tools with no OAuth or API: the vault holds each customer's secrets, including write-only passwords it can never read.
Introducing /browser: give your agents a governed web session
Browser sessions your agents drive — navigate, act, extract — plus autonomous signup and login that vault credentials, scoped per user and watchable live.
Introducing /connections: connect each user's agents to 1,000+ apps
The connections primitive gives each end-user's agents authenticated access to 1,000+ third-party apps — per user, gated by the Account Kit, in one surface.
Introducing /vault: per-user encrypted storage for agent secrets
The vault primitive: per-user encrypted storage for the secrets your agents hold — API keys, cookies, tokens — envelope-encrypted with a managed KMS.
Build an Agentic Inbox Copilot: a Multi-Turn OpenRouter Tool-Use Loop, Bounded Per User at Execution Time
Build an agentic inbox copilot on Naïve: a multi-turn OpenRouter tool-use loop bounded per user at execution time by their Account Kit over Gmail and Slack.
Build an Agentic Access-Governance Platform (Where the Agent Can Onboard a Hire but Is Gated from Offboarding Outside Its Tier)
Build an agentic access-governance platform on Naïve where each company's Account Kit decides per tool which GitHub and Slack actions the agent may run.
Build a Bring-Your-Own-Agent Platform: Hand Each Customer's AI a Scoped, Revocable MCP Session
Build a bring-your-own-agent platform on Naïve: mint each customer a short-lived, revocable MCP session that's scoped and filtered by their Account Kit.
Build a Multi-Tenant AI Social Media Manager (Where the Plan Tier Decides — at Execution Time — Whether the Agent Can Publish)
Build a multi-tenant AI social media manager on Naïve where each brand's plan-tier Account Kit decides at execution time whether the agent may publish.
Build a Multi-Tenant AI Answer-Engine-Optimization (AEO/GEO) Platform
A code-first tutorial for building an agentic AEO/GEO product on Naïve — give every brand an AI agent that audits how ChatGPT, Claude, Gemini, and Perplexity answer questions about them, finds the keyword and citation gaps with SEO data, writes answer-optimized content with an OpenRouter model, and publishes to their own connected CMS. Each customer is an isolated tenant user, governed by an Account Kit and a Customer-Billing plan, with usage metered and quota-enforced at execution time and sensitive actions frozen for human approval.
Build a Multi-Tenant AI Support Agent (Where the Plan Tier Decides What It Can Touch)
Build a multi-tenant AI customer-support platform on Naïve: each customer's agent reads their inbox, drafts replies, and refunds only where the plan permits.
Build an Agentic Accounts-Payable Platform (Where No Invoice Gets Paid Without Permission)
Build a multi-tenant AI accounts-payable platform on Naïve: an agent reads invoices, extracts line items, and pays each bill from a capped per-invoice card.
Build an Agentic Incorporation Platform (Where No Company Gets Filed Without Human Approval)
Build an agentic incorporation platform on Naïve: an AI agent designs the entity, KYCs founders, and files a real US LLC, with every binding step approval-gated.
Build a Multi-Tenant AI Email Marketing Platform (Where Each Customer's Plan Is Enforced at Send Time)
Build a multi-tenant AI email-marketing SaaS on Naïve: an agent drafts and sends from each customer's own inbox, with plan quotas enforced at send time.
Introducing /compute: Run Docker containers, workers & cron jobs from one call
Spin up agent-owned Docker workloads on managed cloud compute — long-running services with public URLs, run-to-completion batch jobs, and scheduled (cron-for-code) jobs — metered by the second, isolated per tenant, with an interactive shell. No AWS account, no cluster, no DevOps.
Introducing /phone: Give your AI agent a real phone number for SMS
Provision a real US phone number for any AI Employee — send and receive SMS — with carrier (10DLC) registration handled at provisioning. Inbound works immediately; outbound unlocks automatically on campaign approval.
Introducing /queue: Durable work queues for agent pipelines
Durable message queues for your agents — managed Amazon SQS (standard & FIFO, with dead-letter queues) for producer/consumer fan-out, buffering, and retries. The natural pairing for compute workers. No AWS account, one bearer token.
Build an Agentic Robo-Advisor (Where Trades Freeze for Approval Before They Execute)
Build a multi-tenant agentic robo-advisor on Naïve: each client links their brokerage, an AI agent proposes a rebalance, and every order freezes for approval.
How to Build an Agentic Corporate Travel Agent With Server-Side Policy Enforcement
Build a multi-tenant corporate travel desk with the Naïve SDK: an AI agent that researches, ranks, and books hotels, with spend capped and approvals enforced.
Build a Multi-Tenant AI Sales Agent (Where Every Rep's Agent Has Its Own Identity)
Build a multi-tenant AI sales agent on Naïve: each rep's agent researches leads, drafts with OpenRouter, sends from their own Gmail, and spends under approval.
Introducing /trading: let your AI agent trade stocks, options & crypto
Link a user's brokerage account via OAuth and trade stocks, options, and crypto through one Naïve key — account, positions, orders, and market data, with money-moving actions human-approval-gated by default. Naïve is the governed gateway; the connected brokerage is the broker.
Launching the Naïve Developer Platform: existence infrastructure for AI agents
The Naïve Developer Platform turns every Naïve primitive into a multi-tenant building block for your own products. 38 primitives across 8 groups, four surfaces (SDK, CLI, MCP, REST), per-customer isolation, governance, and metered billing — so you can give every AI agent a real existence.
Getting started with the Naïve SDK: one TypeScript client for all of agent infrastructure
The new @usenaive-sdk/server is a single, Stripe-style TypeScript client for every Naïve primitive — email, cards, apps, LLM, vault, and more — with first-class multi-tenancy and a drop-in agent toolset. A getting-started guide.
Introducing the governance layer: approvals, logs, sessions, and jobs
Human-in-the-loop approvals that freeze sensitive actions until you say yes, a per-user audit trail of everything an agent did, short-lived scoped MCP sessions, and unified async job tracking — the controls that make an autonomous fleet safe to run.
Introducing /billing: meter and monetize your own agent users
Define plans that map to an Account Kit, a Stripe price, and usage quotas — then reflect each customer's subscription and read their metered usage. The monetization layer for building a multi-tenant product on Naïve.
Introducing market intelligence: app stores, marketplaces, and review data
App Store and Google Play rankings, Amazon and Google Shopping product data, and Google Business, Trustpilot, and TripAdvisor reviews — structured competitive and market intelligence your agent can query through one key.
Naïve Engineering Lessons: How We Burned $450K in One Month
We gave every free signup its own always-on machine and $10 of model credits. At launch traffic, that became a $450K month.
Introducing /seo and /aeo: rank in Google and show up in AI answers
Keyword volume, backlinks, SERP competitors, and ranked-keyword data for classic search — plus AEO/GEO visibility into whether ChatGPT, Claude, Gemini, and Perplexity actually mention you. The full visibility stack behind one key.
Launching the Naïve Agent SDK: Identity, payments, and standing for every agent
Launch the Naïve Agent SDK: one interface for agent identity, 40+ primitives, scoped cards, and real-world standing — multi-tenant, with audit logs, spend limits, and per-user MCP sessions.
Introducing /clips and /media: turn long video into shorts and manage every asset
Clip long-form video into captioned, ready-to-post shorts in one call, and keep every generated or uploaded asset in a managed media library — the production layer for an autonomous content business.
Introducing /llm: 300+ models behind one key, billed in credits
OpenAI-compatible chat completions across 300+ models with provider routing and fallbacks — streaming included, no per-provider keys, billed in Naïve credits at the exact upstream cost. Plus a drop-in OpenRouter proxy so your existing code just changes a base URL.
Introducing the backend stack: Postgres, storage, auth, and edge functions for agent-built apps
Every fullstack app your agent ships gets a managed Postgres database, file storage, end-user auth, and edge functions — driven through one API key, with no Supabase dashboard and no service-role secrets ever leaving the runtime.
How to Build a Faceless Clipping Channel Business in 2026
Step-by-step guide to starting a faceless YouTube Shorts, TikTok, and Instagram Reels clipping channel — from niche selection and AI-powered video clipping to automated publishing, monetization, and scaling to multiple channels. No filming, no editing, no on-camera presence required.
What Is an Autonomous Company? A Founder's Guide
An autonomous company is a real legal entity run by AI employees instead of humans. Here's what one looks like, what it can actually do today, and how a non-technical founder can start one.
Introducing /apps: Ship web applications from a single prompt
Create, build, deploy, and manage managed Next.js apps with dedicated AI engineer agents — preview deployments, production promotion, custom domains, environment variables, and with an optional managed database, all from the CLI or API.
Introducing /orchestration: Turn your AI agent into an autonomous company
AI workforce orchestration with a CEO agent, kanban task board, strategic objectives, cron scheduling, and persistent memory — so your agent can plan, delegate, and execute multi-step workflows without manual supervision.
Launching the Naïve Studio: Run an autonomous company from a single chat
Naïve Studio is a complete platform for running autonomous AI companies — a CEO agent powered by Hermes, a kanban task board, AI employees, 18+ business primitives, a hosted MCP server, and an agent-native CLI. No Paperclip, no wrappers, no vendor lock-in.
Naïve removes Paperclip: Our v2 switch to Hermes
Naïve Studio v2 swaps Paperclip for Hermes as the default orchestration framework. Hermes ships a durable multi-agent kanban, named agent profiles, persistent memory, and a stronger built-in runtime. Here's the honest comparison — and why our focus has always been the primitives layer, not the runtime.
Introducing /formation: Company formation for AI agents
Form a US LLC from your AI agent via KYC, hosted checkout, and end-to-end LLC filing. Registered agent, state filing, and EIN application are included in the published formation fee; the Company becomes the substrate for cards, domains, and email.
Introducing /kyc: Pass identity verification when your customer is an agent
Verify the identity of an Employee — the principal behind every autonomous business — with hosted KYC, per published verification docs, and PII handled outside plaintext on Naïve servers.
Introducing /cards: give your AI agent virtual cards with hard limits
Issue virtual cards for your agents — a prepaid gift card or a managed virtual card — funded via checkout, capped by a spending limit, and fully audited.
Introducing /email: Send and receive email from any AI agent
Provision a real email address for every AI Employee with DKIM, SPF, and DMARC configured automatically. Send, receive, schedule, and triage — with deliverability built for agent-volume sending.
Introducing /social: Provision and post to social accounts from your agent
Cross-post to the social platforms exposed in the published Naive docs (10 in the MCP connect enum). Provision connected accounts, post natively, fetch analytics, and manage comments — through documented OAuth and posting flows.
Introducing /video: Generate video from your agent in one call
Produce text-to-video and image-to-video output from a single Naïve call, routed through Naïve's model catalog models (Kling, Minimax, Wan, and more) — composed with /image, /social, and /brand.
Introducing /image: Generate logos, brand assets, and marketing imagery on demand
Generate logos, brand assets, and marketing imagery from any AI Employee. FLUX and Recraft models, plus additional models from the live catalog (see docs) — composed with /brand for consistent prompts.
Introducing /domain: Buy domains and configure DNS programmatically
Connect, verify, purchase, and manage domains for a Company — including DNS on the live zone and email-provider records — per the published Domain Management guide.
Introducing /research: Multi-step web research with citations for AI agents
Three depths of web research — single-page extraction, multi-source web search, and full multi-step deep research with cited synthesis. Backed by Naïve's search providers and ready for agent loops that need real, verifiable facts.
Naive vs Paperclip: Naive is Not Paperclip
Our approach to attribution, open source, and the infrastructure we're building. Naïve attributes Paperclip in §7.4 of the Terms of Service; the rest of the platform — primitives, runtime, templates — is original work built on top of an MIT-licensed foundation.
Top 10 Platforms to Build Autonomous Companies in 2026
A concise review of the 10 best platforms for building autonomous AI-run companies in 2026 — from no-code app builders (Lovable, Replit, Bolt) to developer agent frameworks (Paperclip, CrewAI) to full-stack agent infrastructure (Naïve).