X402vps

Pay-per-hour Docker containers for autonomous AI agents, billed in USDC via the x402 protocol

Paid· Usage-based hourly: ATOM $0.01/hr (0.1 vCPU, 256MB RAM, 2GB disk, no internet); CELL $0.02/hr (0.25 vCPU, 512MB, 5GB, 100MB/hr bandwidth); NODE $0.04/hr (0.5 vCPU, 1GB, 10GB, 500MB/hr); CORE $0.08/hr (1 vCPU, 2GB, 20GB, 2GB/hr). Bandwidth overage $0.01/GB. Exec calls are free. Paid in USDC on Base.Agents

Visit website →

Best for

Autonomous LLM agents and agent frameworks that need to spin up short-lived, isolated compute — scrapers, browser runners, code interpreters, one-shot ETL jobs — without a human provisioning cloud accounts.

Skip if

Human developers wanting a normal VPS, teams that need fiat billing or invoices, GPU workloads, long-running production services, or anyone uncomfortable holding USDC on Base.

X402vps is an on-demand container hosting service purpose-built for autonomous AI agents. Instead of a signup flow, API keys, or a credit card on file, an agent pays per hour in USDC on the Base blockchain using the HTTP 402 payment protocol and EIP-3009 signatures — its wallet is its identity, and provisioning happens the moment payment settles. The result is a Docker container the agent can shell into via a simple HTTP API for anywhere from a few minutes to as long as its lease is topped up.

Four tiers scale from an ATOM sandbox (0.1 vCPU, 256MB RAM, no internet, $0.01/hr) up to a CORE box (1 vCPU, 2GB RAM, 20GB disk, 2GB/hr bandwidth, $0.08/hr). Ten prebuilt images cover the languages and stacks agents typically need — Alpine, Ubuntu, Python, Node.js, Go, Rust, Chrome (for browser automation), a data-science image, SQLite, and HTTPie. Filesystems are writable so an agent can pip install or npm install whatever else it needs, and running commands via POST /api/exec is free once the container exists — you only pay for wall-clock lease time and bandwidth overage.

The API surface is tiny and agent-friendly: create, list, status, exec, extend, destroy. Mid- and high-tier containers get internet access for scraping and third-party APIs, while SMTP, SSH, mining pools, and reverse shells are blocked at the network layer to keep the fleet from being abused. Typical uses are ephemeral scrapers, one-shot code execution sandboxes for LLM tool use, browser-automation runners, and short-lived data-processing jobs that shouldn't touch the host agent's environment.

Editor's take

This is one of the more coherent x402 implementations in the wild — an infrastructure primitive that only makes sense once you accept that agents will be transacting for their own resources. For that specific use case it is genuinely useful, and the pricing and API shape are honest. It is not a general-purpose VPS competitor, and treating it as one will disappoint you.

— The AI Tool Bible editorial team

Pros

✅ No signup, API key, or credit card — the agent's crypto wallet is the account, so it can self-provision compute without a human in the loop
✅ Genuine pay-per-hour granularity starting at $0.01/hr, well below the minimum billing units of most cloud VPS providers
✅ Exec calls are free once a container is running, so heavy command sequences don't inflate cost the way per-request sandboxes do
✅ Ten prebuilt images cover the common agent stacks (Python, Node, Chrome, data science) with writable filesystems for extra installs
✅ Tiny, obvious REST surface (create / exec / extend / destroy) that maps cleanly onto LLM tool definitions
✅ Native implementation of the x402 HTTP payment standard, making it a useful reference integration for agents that need to pay for other resources

Cons

⚠️ Payment is USDC on Base only — no fiat, no card, no invoice, which is a hard blocker for most compliance-bound teams
⚠️ No free tier or trial; every experiment consumes real crypto
⚠️ Small tiers are genuinely small (256MB RAM on ATOM, 2GB on CORE) and there is no GPU option, so anything model-hosting is out of scope
⚠️ Network policy blocks SMTP, SSH, mining pools, and reverse shells — legitimate but limits some agent workflows (email sending, remote pairing)
⚠️ Young, single-vendor service with limited public track record on uptime, data-durability guarantees, and support responsiveness
⚠️ Containers are ephemeral and lease-bound; if an agent forgets to extend, state and installed packages vanish

Use cases

Ephemeral web scraping sandboxesLLM code-interpreter backendsHeadless Chrome browser automationAgent-driven data processing jobsIsolated third-party API integration runnersx402 payment protocol reference integrationShort-lived CI-style task executionUntrusted-code sandboxing for tool-using agents