> ## Documentation Index
> Fetch the complete documentation index at: https://vetta.sh/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Vetta

> The most efficient managed agent for long-horizon work. The highest intelligence per dollar, by design.

export const Levers = ({items, cols}) => <div style={{
  display: "grid",
  gridTemplateColumns: `repeat(${cols || 2}, minmax(0,1fr))`,
  gap: "12px",
  margin: "1.5rem 0"
}}>
    {items.map((it, i) => <div key={i} style={{
  border: "1px solid rgba(0,0,0,0.10)",
  padding: "16px 18px",
  background: "#ffffff"
}}>
        <div style={{
  fontSize: "13px",
  fontWeight: 500,
  color: "#777777"
}}>{it.k}</div>
        <div style={{
  fontSize: "15px",
  fontWeight: 500,
  margin: "6px 0 4px",
  letterSpacing: "-0.02em",
  color: "#000000"
}}>{it.v}</div>
        <div style={{
  fontSize: "14px",
  color: "#555555",
  lineHeight: 1.5
}}>{it.d}</div>
      </div>)}
  </div>;

**Vetta is a managed agent platform**: you define an [agent](/docs/capabilities/agent) — model, prompt, tools, skills, budget — and run it as [sessions](/docs/concepts/sessions) that work for hours, pause at storage cost, and finish without a person in the loop. You pay per finished task, in real dollars, priced before each call runs.

<Note>
  **Two names, one stack.** Vetta is the autonomous-company infrastructure — the harness, the managed runtime, sessions, budgets, tools. **Naive** is the layer on top: the blueprints and templates a whole company is installed from, and the [studio](https://app.usenaive.ai) you run it in. These pages are the Vetta reference, which is why every page below says Vetta. → [Naive and Vetta](/docs/naive-and-vetta)
</Note>

Everything is built around one number: **the most intelligence per dollar spent** — the same model on different stacks varies several-fold in cost per completed task ([benchmarks](https://usenaive.ai/benchmarks)).

<img src="https://mintcdn.com/vetta/Lby8hzUEjDwI2o6a/images/architecture-overview.png?fit=max&auto=format&n=Lby8hzUEjDwI2o6a&q=85&s=9698e189bb8189f7fdcbdda775751c1f" alt="Vetta — managed agents infrastructure: any model and provider, skills, and tools in; harness, runtime, orchestration, and self-improvement inside; run and interact through the Web Studio, API, CLI, and SDK" width="3200" height="1624" data-path="images/architecture-overview.png" />

Any model, your skills, and your tools go in; Vetta's runtime does the work; and you run and interact through the [Web Studio](https://app.usenaive.ai), the [API](/docs/api/overview), the [CLI](/docs/cli/overview), or the [SDK](/docs/sdk/typescript).

## Create an agent, run a session

<CodeGroup>
  ```typescript TypeScript theme={"system"}
  import { randomUUID } from "node:crypto";
  import { createClient } from "@usenaive-sdk/vetta";

  const vetta = createClient({
    baseUrl: "https://api.vetta.sh",
    apiKey: process.env.VETTA_API_KEY!,
    fetch: globalThis.fetch,
    idempotencyKey: () => randomUUID(),
  });

  const agent = await vetta.agents.create({
    name: "triage",
    model: "zai-org/GLM-5.2-FP8",
    system: "You are a concise triage agent.",
    budget: { cap_micro_usd: 50_000_000, max_task_micro_usd: 5_000_000, period: "month" },
  });

  const session = await vetta.sessions.create({
    agent_id: agent.id,
    message: "Summarize what changed in main overnight.",
  });

  for await (const event of vetta.stream(session.id)) {
    console.log(event.type);
  }
  ```

  ```bash CLI theme={"system"}
  npm i -g @usenaive-sdk/vetta-cli && vetta login --api-key sk_...

  AGENT=$(vetta agent create --name triage --model zai-org/GLM-5.2-FP8 \
    --budget-usd 50 --max-task-usd 5 --budget-period month \
    --system "You are a concise triage agent." | jq -r .id)

  SID=$(vetta session create --agent "$AGENT" \
    --message "Summarize what changed in main overnight." | jq -r .id)

  vetta session stream --session "$SID"
  ```
</CodeGroup>

## The pillars, as objects

Everything you touch maps to primitives you create and version — familiar on day one.

<CardGroup cols={3}>
  <Card title="Agent" icon="robot" href="/docs/capabilities/agent">
    A reusable, versioned config: model, harness, prompt, tools, skills, budget.
  </Card>

  <Card title="Session" icon="comments" href="/docs/concepts/sessions">
    One directly-controlled run of an agent — send input, stream output, interrupt, resume.
  </Card>

  <Card title="Events" icon="wave-square" href="/docs/concepts/events-and-streaming">
    A resumable stream of everything the agent does, keyed by sequence number.
  </Card>

  <Card title="Computer" icon="server" href="/docs/capabilities/computer">
    A disposable sandbox with a filesystem, shell, and optional browser — paused when idle.
  </Card>

  <Card title="Skills" icon="book" href="/docs/capabilities/skills">
    Persisted, versioned expertise loaded with progressive disclosure.
  </Card>

  <Card title="Files" icon="folder" href="/docs/capabilities/files">
    Durable, organization-scoped storage for the artifacts a session produces.
  </Card>
</CardGroup>

## How Vetta earns the highest ROI per dollar

Four independent levers, all cutting dollars per finished task without cutting quality.

<Levers
  cols={2}
  items={[
{ k: "Managed sandbox", v: "Pay only for running time", d: "Every agent gets a disposable micro-VM that pauses between turns — you pay for stored disk, not a machine sitting hot. A fully virtualized V8-isolate mode for near-instant boot is coming soon." },
{ k: "Durable loop", v: "Waiting = storage, not compute", d: "The loop is an alarm, not a call stack: wake, take one turn, commit, sleep. A run that spans hours pays for storage between turns, not a machine sitting hot." },
{ k: "Completion window", v: "One model, three prices", d: "Trade latency for cost per request — immediate, priority, or loose. Same weights, different tariff. Non-urgent work runs on the cheapest lane." },
{ k: "Model + delegation", v: "Right model per step", d: "Pick the model and effort per agent, and delegate sub-tasks to smaller specialists so the long transcript never rides along on every downstream call." }
]}
/>

## Available today & what's next

Available now: agents, sessions, the durable runtime, completion windows, budgets, the computer, skills, files, cron, webhooks, real-USD billing, [identity](/docs/capabilities/identity) (personas, domains, email, SMS, connections, the vault) — plus [web search and fetch, image and video generation](/docs/capabilities/tools#built-in-tools), the [Messages-format model proxy](/docs/api/proxy), [observability](/docs/capabilities/observability), [structured outputs](/docs/capabilities/structured-outputs), [audit logs](/docs/api/audit-logs), and [teams](/docs/capabilities/team) — a coordinator with a version-pinned roster, delegation, and a shared board.

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/docs/quickstart">
    Create an agent, run it, and read the meter in under five minutes.
  </Card>

  <Card title="How Vetta works" icon="layer-group" href="/docs/how-vetta-is-built">
    Harness, runtime, tools — what each layer is, and which one you pick.
  </Card>
</CardGroup>