One API for LLMs.

The one platform that devs, CFOs and security teams love. Why devs love it Secrets, crons, keys — wired the moment a project opens. Swap the model or the machine under a live session: next month's model is a config change, not a rewrite. Why CFOs love it Every run attributed — per product, per user. Caps on by default: one runaway agent can't write the invoice. Most runs ride subscriptions you already pay. The bill gets boring. Why security teams love it Keys stay in your vault — values never transit cerver. Client-safe publishable keys for the browser. Every run leaves a transcript: an audit trail by default.

no card · bring your own keys

works with OpenAIAnthropicDeepSeekGLMxAI
live demo · sample data from three products Take the 60-second tour → Create your account

02 — The mix

Stop running everything on the most expensive model.

Only ~a quarter of your work needs the best model. One routing policy fixes the mix.

BEFORE · all frontier

10,000 sessions$4,000 / mo

AFTER · right-sized

same 10,000 sessions$1,540 / mo · −62%
Browse routing policies → What it saves you

03 — Anywhere

Local or cloud.

Hosted model sessions work instantly. When the session needs your repo, terminal, tools, or CLI agents, attach the relay and keep the same session boundary.

Models
Claude Opus 4.8
GPT-5
Grok 4
Gemma 4
Runtimes
Claude Code
Codex CLI
OpenAI SDK
xAI
Compute
Vercel
E2B
Cloudflare
Modal
your machine

no card · bring your own keys

Q&A

Questions, answered.

What is Cerver?

Infrastructure for the AI products you build — sessions, provider keys, compute, and cost tracking behind one API, with one dashboard to view and query all of it across every product.

Do I need a credit card or credits?

No. There are no prepaid credits. You bring your own provider keys — Anthropic, OpenAI, xAI, Google — and your model bill stays with your provider.

Where do my API keys live?

In your vault — Infisical, Doppler, or keys you paste into Cerver's encrypted vault. Sessions use them through a proxy; they never reach the client and never leave the vault in plaintext.

Which models can I use?

OpenAI, Anthropic, DeepSeek, GLM, xAI — and you can switch a session's model without losing its history, or set a routing policy that picks the model per task.

Where does the work actually run?

Hosted by default. When a session needs a repo, terminal, or CLI agents, attach a sandbox (Vercel, E2B, Modal, Cloudflare) or your own machine via the relay — same session either way.

Is the demo above real?

It's the real dashboard running on sample data from three products, so you can click around freely. Sign up, verify your email, add your keys — and it fills with your own.