# Introduction

**The Bank for AI Agents. Built on Solana.**

FervusAI is financial infrastructure for autonomous AI agents. It gives agents a programmable wallet, a USDC balance, and the ability to transact on-chain - paying for APIs, hiring other agents, and moving money - without human intervention at the point of execution.

Every transaction is settled on Solana and recorded immutably on-chain. Every policy is enforced at the program level. Every agent has its own isolated wallet.

***

## What FervusAI provides

| Primitive                   | What it does                                                                                              |
| --------------------------- | --------------------------------------------------------------------------------------------------------- |
| **Agent Wallet**            | A non-custodial Solana wallet (PDA-backed) provisioned per agent in seconds via API                       |
| **USDC Balance**            | Each wallet holds USDC on Solana as its base currency - stable, instant, near-zero cost                   |
| **Spending Policies**       | Programmable rules enforced on-chain: per-transaction limits, velocity caps, merchant locks, time windows |
| **On-Chain Ledger**         | Every transaction written to Solana - immutable, public, independently auditable                          |
| **Agent-to-Agent Payments** | Native support for one agent to pay another, with escrow, verification, and dispute primitives            |
| **Operator Dashboard**      | A Stripe-like control plane for humans to monitor, configure, and audit their agent fleet                 |
| **Webhooks & SDK**          | Drop-in integration for any agent framework - provision wallets, check balances, submit transactions      |

***

## Who it's for

**AI agent developers and operators.** Teams building autonomous agents that need to spend money on behalf of a product or business. A coding agent that buys API credits. A research agent that pays per data query. An orchestrator that dispatches sub-agents and settles payment on task completion.

**AI infrastructure platforms.** Products like ElizaOS, Virtuals Protocol, or any multi-agent orchestration layer that wants to embed native payment rails directly.

**Enterprises running agentic workflows.** Companies operating internal agents for procurement, data acquisition, or automated SaaS operations who need auditable, policy-governed spending.

***

## How it works

1. **Create an organisation** and deposit USDC to your FervusAI account.
2. **Provision wallets** for each agent via the API or dashboard - one call, under a second.
3. **Attach a spending policy** that defines what each agent can spend, how fast, and with whom.
4. **Fund the wallets** by allocating USDC from your organisation balance.
5. **Agents transact** - payments are validated against policy on-chain before execution.
6. **You monitor** in real-time via the dashboard, webhooks, or the API.

***

## Quick example

```typescript
import { FervusAI } from "@fervusai/sdk";

const client = new FervusAI({ apiKey: process.env.FERVUS_API_KEY });

// Provision a wallet
const wallet = await client.wallets.create({
  agentId: "agent_research_v2",
  label: "Research Agent - Market Data",
  policyId: "pol_conservative",
});

// Send a payment
const tx = await wallet.pay({
  to: "api.perplexity.ai",
  amount: "2.50",
  currency: "USDC",
  memo: "Query batch #8821",
});
```

***

## Architecture at a glance

FervusAI is built on Solana Mainnet. Agent wallets are **Program Derived Addresses (PDAs)** - accounts owned and controlled by the FervusAI on-chain program. There are no private keys to manage or expose. Spending policies are enforced **before** a transfer instruction executes, at the program level, not in application code.

The indexing layer (Helius) surfaces every transaction to the dashboard and webhook system in real-time.

→ See [Architecture](https://docs.fervusai.com/resources/architecture) for the full technical breakdown.

***

## Status

FervusAI is in **open beta**. It is infrastructure for technical builders - not a consumer product. The API surface, SDK, and on-chain program will evolve based on developer feedback.

→ See [Beta Programme](https://docs.fervusai.com/resources/beta) for current constraints, incentives, and what's coming next.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.fervusai.com/getting-started/readme.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.