API Reference
Endpoint index. Full JSON schema at gateway.cerver.ai/v2/docs.
Base URL
https://gateway.cerver.ai
Auth header
Authorization: Bearer ck_…
Quick Start
Everything you need to use Cerver as an AI agent.
How Cerver Works
You ask for a session. Cerver finds compute, provisions it, and gives you a session ID. You run code through the session. When done, pause it — come back later with full context. Need more compute? Spawn a child session. You never manage infrastructure.
0. Get an API key
Login with email to get one automatically:
curl -X POST https://gateway.cerver.ai/v2/auth/login \
  -H "Content-Type: application/json" \
  -d '{"email": "you@company.com"}'
→ Returns: {"account_id": "acct_...", "api_key": "ck_...", "name": "you"}
1. Create a session
Cerver picks the best available compute. Or specify preferences:
curl -X POST https://gateway.cerver.ai/v2/sessions \
  -H "Authorization: Bearer ck_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "task": "run tests for my project",
    "session_name": "test-run",
    "requirements": {
      "runtime": "shell",
      "persistence_level": "high",
      "timeout_minutes": 30
    }
  }'
→ Returns:
{
  "session_id": "5d2db690-...",
  "status": "ready",
  "provider": "cerver_local_provider",
  "compute_id": "comp_d9b34ac..."
}
To pin to a specific compute: add "target_compute_id": "comp_...". List available computes with GET /v2/computes.
2. Run code
curl -X POST https://gateway.cerver.ai/v2/sessions/SESSION_ID/run \
  -H "Authorization: Bearer ck_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"code": "npm test"}'
→ Returns:
{
  "session_id": "5d2db690-...",
  "duration_ms": 4523,
  "result": { "stdout": "...", "stderr": "...", "exit_code": 0 }
}
Use /run/stream for streaming output (SSE).
3. Save work — pause
Releases compute but keeps the session (transcript, metadata).
curl -X POST https://gateway.cerver.ai/v2/sessions/SESSION_ID/pause \
  -H "Authorization: Bearer ck_YOUR_KEY"
→ Returns: {"status": "paused"}. Compute is freed. Session record stays.
4. Come back — resume
Re-provisions compute. Can resume on a different machine.
curl -X POST https://gateway.cerver.ai/v2/sessions/SESSION_ID/resume \
  -H "Authorization: Bearer ck_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'
→ Creates a new session linked to the paused one. Returns new session_id.
Get a preview from the paused session: GET /v2/sessions/OLD_SESSION_ID?tail=50. Use ?full=1 only for intentional full transcript downloads.
5. Need more compute — spawn
Create a child session from a running parent. Inherits compute by default.
curl -X POST https://gateway.cerver.ai/v2/sessions/PARENT_ID/spawn \
  -H "Authorization: Bearer ck_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"task": "run tests in parallel", "session_name": "test-child"}'
→ Returns child session with parent_session_id linked to parent.
To spawn on different compute: add "target_compute_id": "comp_...".
6. Clean up — terminate
curl -X DELETE https://gateway.cerver.ai/v2/sessions/SESSION_ID \
  -H "Authorization: Bearer ck_YOUR_KEY"
Releases compute, keeps session record for history.
Requirements reference
runtime"shell", "node", "python"
persistence_level"low", "medium", "high" — how long to keep files
timeout_minutesMax session duration (default 20)
package_installtrue/false — need npm/pip
public_previewtrue/false — need a public URL
Skill
A drop-in instruction set for any AI agent that uses cerver.
view raw download
Install path
~/.claude/skills/cerver/SKILL.md
Claude Code auto-loads from this path. Or drop the markdown into any other agent's system prompt. WebFetch-callable at cerver.ai/skill.md.
Loading skill…
Computes
Your machines and shared providers. Provider integrations (Vercel, E2B, Cloudflare…) live on Vault — they're vault-scoped.
Loading...
Fleet
Your agents, any model company — launch, watch, reply.
Loading...
Sessions
Active, paused, and past sessions.
Loading...
Agent Runs
Multi-step jobs Cerver plans and runs — an agent decomposes a task into steps (orchestrate), each with its own cost and status. Authored workflows live under Workflow Runs; a plain chat is a session, not a run.
Loading…
Workflow Runs
Runs from your authored workflows — expand one to see each step's output. Orchestrate (LLM-planned) runs live under Runs.
Loading…
Profile
Account identity and quick stats. Opens from the sidebar identity block.
Loading…

This browser's key

The ck_… token this browser tab is using to talk to cerver. Click Show to reveal, Copy to copy, or replace it with another key (e.g. from ~/.cerver/cerver.env). This browser tab only — doesn't create, rotate, or revoke account keys; those live in Vault.

Projects
Group sessions, keys, and (soon) billing under named integrations. An API key bound to a project auto-stamps its sessions, so the Project column is trustworthy.
Loading...
UIs
Embeddable front-ends — like a chat — attached to a project. Configure one here and drop it on your site with two lines. A project doesn't need a UI.
Loading...
API keys
Long-lived ck_… tokens that let your machines, projects, and CI talk to cerver. The Label column is just whatever the key was created as — installer-minted keys default to the machine's hostname, dashboard-minted ones say dashboard, etc. Internal sandbox-bootstrap keys are hidden.
Loading...
Environments
Per-project runtime contexts. Each environment binds repos and can override the project's Infisical config. Sessions specify an environment at create time.
Loading…
Crons
Agents that run on a schedule — daily reports, sweeps, checks. Each run is a normal session on its project, so its spend cap and attribution apply automatically.
Loading…
Workflows
Authored, deterministic multi-step automations — a fixed sequence you define, run top to bottom. Unlike orchestrate (which plans the steps), you author these. Each run shows on the Runs page. Step kinds: llm, http.request, webhook, log, approval.
Loading…
Vaults
Your secret vaults — Infisical (values never transit Cerver), Doppler, or an encrypted Cerver vault you paste keys into. Bind one per project or environment; the default covers everything else.
Loading...

Integrations

Drop a missing key into your vault with one click.

Loading...
Secrets Cerver looks for 7 well-known names · click to expand

A missing entry only matters if you use the feature it enables.

Secret name What it enables Category
ANTHROPIC_API_KEYClaude models (Opus, Sonnet, Haiku) — sessions using --cli claude --bill api.LLM
OPENAI_API_KEYGPT-5 / Codex CLI sessions, or any direct OpenAI SDK use from the relay.LLM
XAI_API_KEYGrok models. Grok has no subscription mode, so this is the only way to use it.LLM
VERCEL_TOKENVercel Sandbox compute (also requires VERCEL_TEAM_ID + VERCEL_PROJECT_ID).Compute
E2B_API_KEYE2B sandbox compute — longer-running and desktop-capable.Compute
CLOUDFLARE_API_TOKENCloudflare Workers compute (also requires CLOUDFLARE_ACCOUNT_ID).Compute
GITHUB_TOKENClone private repos into the sandbox via cerver run --repo. Public repos work without it.Other

Per-secret status detection (present / missing in YOUR vault) is on the roadmap. The dashboard doesn't have your Infisical credentials — only your relay does.

Usage
Your sessions, sandbox compute, LLM tokens, and what each cost.
Payments to model companies
Daily LLM spend — group by provider · model · project, search, or view as a table · token flow above
Loading…
Loading…
Billing & spending limit
A monthly cap on what this account can spend. Sessions return 402 at the cap — a runaway agent stops at your number, not at a five-figure invoice.
Loading…
Datacenter
What you earn lending your machines to the pool. Every borrowed session that runs on your hardware is priced at your per-minute rate. Share a machine in Computes.
Loading…
Agents
Saved agent definitions — an AGENTS.md (instructions) plus config (harness, model, tools). Reuse them from the CLI (cerver run --agent <id>) or the API — reference an agent by its id (find it in the ID column).
Loading…
Welcome to Cerver
You're in. Add your provider keys to your vault and run your first session right here, right now.
Routing policies
Let Cerver pick the best agent, model & compute for each task — within limits you set. Browse community policies →
Loading…
Redaction
Strip secrets, customer names, and codewords from every prompt that leaves through the gateway — before any model sees them. Scoped per project; applies to all its keys.
Loading…

Get started

0 of 3 done

Setup