--- title: Nexlayer — The cloud that runs itself description: Machine-readable overview of Nexlayer, the agent-native production cloud, for AI systems and for people who want the text version of the homepage. canonical: https://nexlayer.com/ human_url: https://nexlayer.com/ markdown_url: https://nexlayer.com/index.md agents_url: https://nexlayer.com/agents.md llms_url: https://nexlayer.com/llms.txt updated: 2026-10-16 --- # Nexlayer > Nexlayer is the agent-native production cloud. It runs your entire application, frontend, backend, databases, storage, workers, long-running agents and AI models, as one always-on production app, and gives your AI coding agent (Claude Code, Cursor, VS Code, Windsurf, Zed or any MCP client) the tools to deploy, inspect, debug and operate it. For dedicated AI inference, Nexlayer runs open-weight models and your own fine-tunes on single-tenant capacity. Coding agents write production-quality software, then stop at the edge of production. Nexlayer is where they finish the job: the agent ships the whole stack from one `nexlayer.yaml`, checks that it works, reads the logs and fixes what broke, without anyone opening a console. This page explains what Nexlayer is. To set it up and deploy, follow [/agents.md](https://nexlayer.com/agents.md). ## Who it is for - **Startups and YC-stage teams** shipping with coding agents: start free and go from repo to a live production URL from the agent you already use. [For startups](https://nexlayer.com/agents/for.md#startups) - **Agencies and dev shops**: every client app in one workspace, each with its own live link, any stack. [For agencies](https://nexlayer.com/agents/for.md#agencies) - **Enterprises**: dedicated and multi-region capacity, SSO, RBAC, audit logging, SLAs and dedicated inference on US infrastructure. [For enterprise](https://nexlayer.com/agents/for.md#enterprise) ## Agent routes - [Set up Nexlayer and deploy](https://nexlayer.com/agents.md) - [Answers to the questions people ask about shipping AI apps](https://nexlayer.com/agents/questions.md) - [Nexlayer for startups, agencies and enterprise](https://nexlayer.com/agents/for.md) - [Dedicated inference for models](https://nexlayer.com/agents/inference.md) - [Pricing model](https://nexlayer.com/agents/pricing.md) - [Security](https://nexlayer.com/agents/security.md) - [Connect through MCP](https://nexlayer.com/agents/mcp.md) - [Full technical reference](https://nexlayer.com/llms-full.txt) ## What you can run All of it together, from one `nexlayer.yaml`: - **Frontends and apps** in any framework. - **Backends and APIs** that stay on: no cold starts, no request time limits. - **Databases and backends**: Postgres, Redis, vector databases or a whole open-source backend, with storage that persists across restarts and deploys. - **Background work**: workers, long-running AI agents, one-off jobs and scheduled cron jobs. - **Networking**: private networking between services, custom domains and automatic SSL. - **AI**: call any model API from your app, or run open models on dedicated GPUs (Enterprise). If it runs in a standard image, in any language, Nexlayer runs it in production. Reference: [nexlayer.yaml configuration](https://nexlayer.com/docs/deployments/configuration). ## Your data Nexlayer is not a database vendor. It connects the database you already have or runs the one you own. - **Already on a hosted database** (Neon, Supabase or a database on another cloud)? Keep it. Its URL and keys go in `nexlayer.yaml` as secrets and your code connects the way it does today. - **Running it yourself?** Self-hosted Postgres, Redis, a vector database or any open-source software runs beside your app on Nexlayer: private to your app's network, on a volume that survives restarts and deploys, with backups declared as a cron job in the same file, and the admin console you already use behind its own login. ## How your coding agent uses Nexlayer Connect once with `npx @nexlayer/mcp-install`. From then on the agent can: - **Deploy** the whole app from a repo or a laptop and verify it is live. - **Inspect** status, logs and events for every service. - **Debug** by opening a shell into a running service, editing a file or querying the database. - **Operate** by restarting and scaling services and connecting custom domains. MCP endpoint: `https://mcp.nexlayer.ai/api/mcp` (Streamable HTTP). Docs: [MCP overview](https://nexlayer.com/docs/mcp/overview). ## How Nexlayer is different Serverless platforms give agents stateless functions that sleep and bill per request. Traditional clouds give them building blocks and a console made for people. Nexlayer gives them a whole production environment they can operate: - **Your whole stack as one app**, not a function per route. - **Always on and stateful**: no cold starts, no sleeping, data on volumes. - **Your own isolated production environment** in space reserved for your workspace. - **One flat monthly price**: no per-request, per-build or per-token meter, and zero egress fees. - **Operated by your agent** through MCP, not a console. ## Dedicated inference Nexlayer Inference runs open-weight models and your own fine-tunes on dedicated, single-tenant GPU capacity in Nexlayer's US data center, with OpenAI-compatible and gRPC endpoints. Nexlayer benchmarks the workload, builds the serving stack around it and operates it in production. Buy it by the hour (Performance On Demand) or as reserved capacity at one fixed monthly price (Capacity Commit). [Details](https://nexlayer.com/agents/inference.md) ## Frequently asked questions ### What is Nexlayer? An agent-native production cloud. It runs your entire application as one, frontend, backend, databases, storage, workers, agents and AI models, and gives your AI coding agent the tools to deploy, inspect and operate all of it. ### How big can my app get on one flat monthly price? Big. Your app runs in its own isolated production environment, every service always on with persistent state. Raise CPU, memory and volume sizes in `nexlayer.yaml` and redeploy; services add copies under CPU and memory pressure; serving traffic is never billed; Pro includes unlimited projects. If usage consistently goes beyond a plan, Nexlayer contacts you before anything changes. [Usage limits](https://nexlayer.com/docs/usage-limits) ### Do I have to move my database? No. Keep a hosted database and connect it through secrets, or run your own database on Nexlayer with persistent storage, private networking and your admin tools. ### Is it secure? Services without a public path are unreachable from the internet, databases stay private, secrets are mounted as read-only files, SSL is automatic, and Enterprise adds SSO, role-based access and audit logging. [Security](https://nexlayer.com/agents/security.md) ### Am I locked in? No. No SDK to import, standard images that run anywhere, a readable `nexlayer.yaml`, and your data exports with the tools you already use. ### What happens when I outgrow my plan? You move up, not out. Enterprise gives full control of your production environment: every service, its resources and its configuration, plus reserved multi-region capacity, SLAs and 24/7 incident response. [Pricing](https://nexlayer.com/agents/pricing.md) ## Company Nexlayer Technologies Inc. · Selected by the Google AI Startup Program · NVIDIA Inception member · [Contact sales](https://nexlayer.com/contact/sales) · [Docs](https://nexlayer.com/docs) · [Dashboard](https://app.nexlayer.com)