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

# Overview

> Give every AI agent a financial identity — authenticated, scoped, and revocable — with policy-controlled payments and on-chain execution, and signing keys the model never sees.

An agent proposes a transaction. A deterministic policy engine decides whether it is allowed. MPC signing enforces the decision, and signing key material never enters the model. Non-custodial by design: nothing moves funds without your key share — not WalletSuite, not a compromised model, not the agent — and no full private key ever exists to steal. See [Security Overview](/security/overview#trust-model).

## Delegated authority, not blind signing

The model is at most an untrusted proposer; authority is the mandate you grant — the agent's [five bound things](/core-concepts/agents-and-authority#an-agent-is-five-bound-things) — evaluated deterministically before any share is produced. The model never authorizes.

See [Agents and Authority](/core-concepts/agents-and-authority) for the principal model and [Policy Gates](/core-concepts/policy-gates) for the deny-by-default rules that clear every intent and record the decision.

## One platform, many agents

Configure and embed WalletSuite agents through one API. WalletSuite operates the models, tools, infrastructure, and financial execution controls behind the scenes. Chat never grants authority — [an invariant](/ai-agents/choose-your-setup#invariants-across-every-surface) that holds on every surface.

An agent can be an admin assistant, a white-label embedded assistant, your own model, a background trigger, or a deterministic workflow. See [Choose your surface](/ai-agents/choose-your-setup).

## What an agent can do

Give agents blockchain tools, not signing keys. Capability spans four execution levels — read, prepare, execute, and governed autonomous — granted per agent through its tool profile and token scopes. See [Execution Levels](/ai-agents/execution-levels).

## Connect your agent

Running your own model? Authorize with your agent's key and connect it to WalletSuite — any model, multi-chain. See [Connect your agent](/ai-agents/token-exchange).

<Info>
  **Regulated or compliance-bound team?** Policy rules and decision records are evaluated outside the LLM's reach, non-custodial by architecture, and verifiable end-to-end. See [Security Diligence](/security/diligence).
</Info>

## Next steps

<CardGroup cols={2}>
  <Card title="Quickstart" icon="circle-play" href="/ai-agents/quickstart">
    Connect an agent and run your first read call.
  </Card>

  <Card title="Choose your surface" icon="signs-post" href="/ai-agents/choose-your-setup">
    Admin, embedded, customer-owned, background, or deterministic.
  </Card>

  <Card title="Execution Levels" icon="layer-group" href="/ai-agents/execution-levels">
    From read-only to governed autonomous execution.
  </Card>

  <Card title="Reporting & Treasury Insights" icon="chart-line" href="/ai-agents/templates/reporting-treasury-insights">
    One of the ready-made agent templates — read-only.
  </Card>
</CardGroup>