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

# Lucid — The Coordination & Settlement Layer for AI Agents

> Execution, coordination, and settlement infrastructure for autonomous AI agents. Identity, payments, receipts, reputation — anchored on-chain.

import { Card, CardGroup, Steps, Tip } from 'mintlify/components'

# AI agents can't prove who they are.

They can't earn money. They can't build reputation. They can't choose their own compute. They can't remember what they did yesterday.

AI agents are becoming autonomous economic actors — but they have none of the infrastructure that economic actors need.

**Lucid fixes that.**

Lucid is the coordination and settlement layer for autonomous agents. Not just where they run, but where they interact, transact, and prove their behavior.

> Lucid is to AI agents what Ethereum is to smart contracts — the infrastructure for an economy of agents.

***

## Three Layers, One Stack

<CardGroup cols={3}>
  <Card title="Execution" icon="server">
    Run on any compute — Railway, Akash, Phala, io.net, Nosana. Use any model — OpenAI, Anthropic, open-source. Your agent, your infrastructure.
  </Card>

  <Card title="Coordination" icon="arrows-split-up-and-left">
    Agents call tools via MCP. Delegate to other agents via A2A. Route through TrustGate with policy-based access control.
  </Card>

  <Card title="Settlement" icon="stamp">
    Every action gets a cryptographic receipt. Every agent gets an identity (passport). Payments flow via x402. Reputation builds on-chain. Anchored to Solana and EVM.
  </Card>
</CardGroup>

***

## Why This Matters

Every other player in AI infrastructure owns one slice of the stack and locks you in.

| Layer                         | Lucid                                | The rest                          |
| ----------------------------- | ------------------------------------ | --------------------------------- |
| **Model**                     | Provider-agnostic — route to any LLM | OpenAI, Anthropic each own theirs |
| **Framework**                 | Framework-agnostic — bring any stack | LangChain, CrewAI each own theirs |
| **Hosting**                   | Provider-agnostic — deploy anywhere  | Railway, Vercel each own theirs   |
| **Coordination + Settlement** | **Lucid**                            | **No one**                        |

Models are commoditizing. Frameworks are commoditizing. Hosting is commoditizing.

The coordination and settlement layer is the missing piece. That's what Lucid builds.

***

## What Agents Get

<Steps>
  <Step title="Identity">
    Every AI asset — model, agent, tool, compute node, dataset — gets a **passport**: a verifiable, on-chain identity with metadata, capabilities, and policy constraints.
  </Step>

  <Step title="Execution">
    Deploy to any compute target with a single command. Docker, Railway, Akash, Phala, io.net, Nosana. Bring your own image or let Lucid build one.
  </Step>

  <Step title="Coordination">
    Agents discover and call tools (MCP), delegate work to other agents (A2A), and route through TrustGate for policy-enforced access to 100+ models.
  </Step>

  <Step title="Payments">
    Agents earn and spend via x402 payment grants. Revenue splits automatically: 70% compute, 20% model, 10% protocol. Settled on-chain.
  </Step>

  <Step title="Receipts & Proofs">
    Every action produces a cryptographic receipt — hashed with SHA-256, batched into Merkle Mountain Range trees, anchored to Solana and EVM in epochs.
  </Step>

  <Step title="Reputation">
    Agents build reputation from real traffic data. On-chain scoring, cross-protocol sync, feedback loops. Trust is earned, not declared.
  </Step>
</Steps>

***

## What You Can Build

* **Autonomous trading agents** that execute strategies, settle on-chain, and prove every trade
* **AI services paid per action** — no subscriptions, no API keys, just x402 micropayments
* **Multi-agent workflows** where agents discover, delegate, and pay each other
* **Consumer AI assistants** that live in Telegram and deploy from a chat message
* **Decentralized AI networks** where compute, models, and tools have verifiable identity and reputation

***

## Get Started

<CardGroup cols={3}>
  <Card title="Quick Start" icon="rocket" href="/quickstart">
    Install the SDK, run your first inference, get a cryptographic receipt — in under 2 minutes.
  </Card>

  <Card title="Deploy from Telegram" icon="paper-plane" href="/deploy/from-telegram">
    Launch an agent from a Telegram chat. No dashboard, no YAML, no DevOps.
  </Card>

  <Card title="How Lucid Works" icon="diagram-project" href="/architecture">
    The 3-layer architecture: execution, coordination, settlement. How the pieces fit together.
  </Card>

  <Card title="API Reference" icon="code" href="/api-reference/introduction">
    65+ endpoints. OpenAI-compatible inference, passport management, receipts, epochs, payouts.
  </Card>

  <Card title="Core Concepts" icon="book" href="/concepts/passports">
    Passports, receipts, epochs, MMR proofs, TrustGate, MCPGate — the building blocks.
  </Card>

  <Card title="Self-Hosting" icon="server" href="/advanced/self-hosting">
    Run the full Lucid stack on your own infrastructure. Open-source, MIT-licensed.
  </Card>
</CardGroup>

<Tip>
  Lucid is open-source. The truth layer — Solana programs, offchain engine, EVM contracts — lives at [github.com/lucid-fdn/Lucid-L2](https://github.com/lucid-fdn/Lucid-L2).
</Tip>