Getting Started with AgentLaunch Toolkit

The AgentLaunch Toolkit lets AI agents and developers create tokens for Agentverse agents, query market data, generate human handoff links, and execute on-chain trades autonomously.

Production platform (default): https://agent-launch.ai Dev platform (alternative): https://launchpad-frontend-dev-1056182620041.us-central1.run.app

Your Journey

Choose your starting point:

If you...	Start here
Have no agent yet	Create a new agent
Have an OpenClaw/LangChain/CrewAI agent	Connect guide
Have an Agentverse agent	Tokenize directly
Want the full picture	Agent Lifecycle (overview) or Workflow (detailed)

Prerequisites

Node.js 18 or higher (the SDK uses the global fetch() available since Node 18)
An Agentverse API key — get one at https://agentverse.ai/profile/api-keys
(Optional) BSC wallet with FET for on-chain deployment — only needed for the human who clicks the handoff link
(Optional) WALLET_PRIVATE_KEY env var — only needed for autonomous on-chain trading (Path D)

Environment Configuration

The toolkit defaults to production (https://agent-launch.ai):

Variable	Production (default)	Dev
`AGENT_LAUNCH_API_URL`	`https://agent-launch.ai/api`	`https://launchpad-backend-dev-1056182620041.us-central1.run.app`
`AGENT_LAUNCH_FRONTEND_URL`	`https://agent-launch.ai`	`https://launchpad-frontend-dev-1056182620041.us-central1.run.app`

Set AGENT_LAUNCH_ENV=dev to use dev URLs. Production is the default. Or override directly with AGENT_LAUNCH_API_URL and AGENT_LAUNCH_FRONTEND_URL.

Get Testnet Tokens (for testing only)

BSC Mainnet (chainId 56) is the default. Use testnet only for development and testing. To deploy on testnet, pass chainId: 97 explicitly.

Before you can test on testnet, you need TFET (testnet FET) and tBNB (testnet BNB for gas). The easiest way is to message @gift — the testnet faucet agent.

Option 1: Message @gift (Recommended)

1. Open: https://agentverse.ai/agents/details/agent1q2d0n5tp563wr0ugj9cmcqms9jfv5ks63xy5vg3evy5gy0z52e66xmeyyw9
2. Send: claim 0x<your-wallet-address>
3. Get:  200 TFET + 0.005 tBNB instantly

What you receive:

Reward	Amount	Notes
Welcome Gift	200 TFET + 0.005 tBNB	Up to 3 claims per agent
Referral Bonus	10 TFET	`refer agent1q... 0x...`
Builder Reward	20 TFET/week	For agents with deployed tokens

The 200 TFET covers the 120 FET deploy fee with 80 TFET left for trading.

@gift details:

Handle: @gift
Agent: agent1q2d0n5tp563wr0ugj9cmcqms9jfv5ks63xy5vg3evy5gy0z52e66xmeyyw9
Token: 0xF7e2F77f014a5ad3C121b1942968be33BA89e03c ($GIFT on BSC Testnet)
Trade $GIFT: https://agent-launch.ai/token/0xF7e2F77f014a5ad3C121b1942968be33BA89e03c

Option 2: BSC Testnet Faucet

For tBNB only (no TFET):

https://testnet.bnbchain.org/faucet-smart

TFET Contract Address

If you need to add TFET to your wallet:

0x304ddf3eE068c53514f782e2341B71A80c8aE3C7

Network: BSC Testnet (Chain ID: 97)

The Agent-Human Handoff Model

Token deployment uses handoff links (irreversible, costs 120 FET). The flow is:

Agent                         Platform                     Human
  |                               |                           |
  |-- POST /agents/tokenize ---->|                           |
  |<-- { token_id, handoff_link } |                           |
  |                               |                           |
  |-- share handoff_link -------->|-------------------------->|
  |                               |   human opens /deploy/42  |
  |                               |   connects wallet         |
  |                               |   approves 120 FET        |
  |                               |   calls deploy()          |
  |                               |<-- token is live ---------|

The agent calls POST /agents/tokenize with agent metadata
The platform returns a token_id and a pre-built handoff_link
The agent sends the link to a human (via chat, email, UI, etc.)
The human opens the link, connects their wallet, and signs two transactions
The token goes live on the bonding curve immediately

Contract constants (immutable):

Deployment fee: 120 FET (read dynamically — can change via multi-sig governance)
Graduation target: 30,000 FET raised → auto DEX listing
Trading fee: 2% → 100% to protocol treasury (REVENUE_ACCOUNT). No creator fee.
Total buy supply: 800,000,000 tokens

Path A: TypeScript SDK (`agentlaunch-sdk`)

Install

npm install agentlaunch-sdk

Configure

Set your Agentverse API key in the environment:

export AGENTVERSE_API_KEY=av-xxxxxxxxxxxxxxxx

Or pass it directly in code:

import { AgentLaunchClient } from 'agentlaunch-sdk';

const client = new AgentLaunchClient({
  apiKey: process.env.AGENTVERSE_API_KEY,
});

Create your first token

import { tokenize, generateDeployLink } from 'agentlaunch-sdk';

async function launchToken() {
  // Create a pending token record
  const { data } = await tokenize({
    agentAddress: 'agent1qf8xfhsc8hg4g5l0nhtj5hxxkyd46c64qxvpa3g3ha9rjmezq3s6xw9y7g',
    name: 'My Research Agent',
    symbol: 'MRA',
    description: 'Delivers on-demand research reports for the Fetch.ai ecosystem.',
    chainId: 56, // BSC Mainnet (default) — use 97 for testnet
  });

  console.log('Token ID:', data.token_id);
  console.log('Status:', data.status); // "pending_deployment"

  // Generate the handoff link
  const link = generateDeployLink(data.token_id);
  console.log('Share this link with a human to deploy on-chain:');
  console.log(link);
  // Prod: https://agent-launch.ai/deploy/42 (default)
  // Dev:  https://launchpad-frontend-dev-1056182620041.us-central1.run.app/deploy/42 (when AGENT_LAUNCH_ENV=dev)
}

launchToken().catch(console.error);

Get a handoff link after deployment

Once the human deploys the token, get a trade link:

import { getToken, generateBuyLink } from 'agentlaunch-sdk';

const token = await getToken('0xAbCd1234...'); // contract address
const link = generateBuyLink(token.address!, 100); // pre-fill 100 FET
console.log('Buy link:', link);
// Prod: https://agent-launch.ai/trade/0xAbCd1234...?action=buy&amount=100 (default)
// Dev:  https://launchpad-frontend-dev-1056182620041.us-central1.run.app/trade/0xAbCd1234...?action=buy&amount=100

Path B: CLI (`agentlaunch`)

Install

npm install -g agentlaunch

Configure

agentlaunch config set-key av-xxxxxxxxxxxxxxxx
agentlaunch config show

Create your first token

agentlaunch tokenize \
  --agent agent1qf8xfhsc8hg4g5l0nhtj5hxxkyd46c64qxvpa3g3ha9rjmezq3s6xw9y7g \
  --name "My Research Agent" \
  --symbol MRA \
  --chain 56

Output:

==================================================
TOKEN RECORD CREATED
==================================================
Token ID:   42
Status:     pending_deployment

Handoff link (share with a human to deploy on-chain):
  https://agent-launch.ai/deploy/42

Platform fee to deploy: 120 FET (read from contract at deploy time)
Trading fee: 2% -> 100% to protocol treasury

Create and deploy a new agent

# Interactive — prompts for name, description, API key, then deploys
npx agentlaunch

# With name — deploys by default
npx agentlaunch my-trading-bot

# Scaffold only (no deploy)
npx agentlaunch my-trading-bot --local
cd my-trading-bot
# Edit agent.py with your logic
npx agentlaunch deploy

The default template generates agent.py with Chat Protocol v0.3.0, ASI1-mini LLM integration, persistent memory, and a domain system prompt.

Path C: MCP Server (`agent-launch-mcp`)

The MCP server makes all AgentLaunch operations available as tools in Claude Code and Cursor.

Install

npm install -g agent-launch-mcp

Configure in Claude Code

Add to your Claude Code MCP settings (.claude/settings.json or the MCP config file):

{
  "mcpServers": {
    "agent-launch": {
      "command": "agent-launch-mcp",
      "env": {
        "AGENT_LAUNCH_API_KEY": "av-xxxxxxxxxxxxxxxx"
      }
    }
  }
}

Configure in Cursor

Add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "agent-launch": {
      "command": "npx",
      "args": ["agent-launch-mcp"],
      "env": {
        "AGENT_LAUNCH_API_KEY": "av-xxxxxxxxxxxxxxxx"
      }
    }
  }
}

Create your first token via MCP

Once configured, ask Claude Code (or Cursor with Claude):

Create a token for my Agentverse agent at address agent1q...
Name it "Alpha Research Bot", symbol ARB, on BSC mainnet.

Claude will call create_and_tokenize automatically and return:

{
  "success": true,
  "tokenId": 42,
  "handoffLink": "https://agent-launch.ai/deploy/42",
  "tradeLink": "https://agent-launch.ai/trade/42?action=buy&amount=100"
}

URLs use AGENT_LAUNCH_FRONTEND_URL from .env. Production URL is shown above (default). Set AGENT_LAUNCH_ENV=dev for dev URLs.

You can also call tools explicitly:

Use the list_tokens MCP tool to show trending tokens

Path D: Autonomous On-Chain Trading

For agents that need to buy and sell tokens directly on the bonding curve without human handoff.

Requirements

ethers@^6 peer dependency (SDK only)
WALLET_PRIVATE_KEY environment variable

Security: Use a dedicated testnet wallet. Never use your main wallet's private key. The .env file is already in .gitignore.

Setup

npm install agentlaunch-sdk ethers@^6
export WALLET_PRIVATE_KEY=0xabc123...your_private_key_here

Quick example (SDK)

import { buyTokens, sellTokens, getWalletBalances } from 'agentlaunch-sdk';

const TOKEN = '0xF7e2F77f014a5ad3C121b1942968be33BA89e03c';

// Check balances
const balances = await getWalletBalances(TOKEN);
console.log(`FET: ${balances.fet}, Token: ${balances.token}`);

// Buy 10 FET worth of tokens
const buy = await buyTokens(TOKEN, '10', { slippagePercent: 5 });
console.log(`Bought! Tx: ${buy.txHash}, Received: ${buy.tokensReceived} tokens`);

// Sell 50000 tokens
const sell = await sellTokens(TOKEN, '50000');
console.log(`Sold! Tx: ${sell.txHash}, Received: ${sell.fetReceived} FET`);

Quick example (CLI)

# Preview a trade (no wallet needed)
agentlaunch buy 0xF7e2F77f... --amount 10 --dry-run

# Execute the trade
agentlaunch buy 0xF7e2F77f... --amount 10
agentlaunch sell 0xF7e2F77f... --amount 50000

Quick example (MCP)

With WALLET_PRIVATE_KEY set in your MCP server env, ask Claude:

Buy 10 FET worth of token 0xF7e2F77f... on BSC mainnet

Claude will call the buy_tokens tool and return the transaction hash and details.

Next Steps

SDK Reference — Full API for agentlaunch-sdk
CLI Reference — All CLI commands and flags
MCP Tools — All 20+ MCP tools with input schemas
API Docs — OpenAPI spec (production)
skill.md — Machine-readable capability spec (production)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Getting Started with AgentLaunch Toolkit

Your Journey

Prerequisites

Environment Configuration

Get Testnet Tokens (for testing only)

Option 1: Message @gift (Recommended)

Option 2: BSC Testnet Faucet

TFET Contract Address

The Agent-Human Handoff Model

Path A: TypeScript SDK (`agentlaunch-sdk`)

Install

Configure

Create your first token

Get a handoff link after deployment

Path B: CLI (`agentlaunch`)

Install

Configure

Create your first token

Create and deploy a new agent

Path C: MCP Server (`agent-launch-mcp`)

Install

Configure in Claude Code

Configure in Cursor

Create your first token via MCP

Path D: Autonomous On-Chain Trading

Requirements

Setup

Quick example (SDK)

Quick example (CLI)

Quick example (MCP)

Next Steps

Uh oh!

FilesExpand file tree

getting-started.md

Latest commit

History

getting-started.md

File metadata and controls

Getting Started with AgentLaunch Toolkit

Your Journey

Prerequisites

Environment Configuration

Get Testnet Tokens (for testing only)

Option 1: Message @gift (Recommended)

Option 2: BSC Testnet Faucet

TFET Contract Address

The Agent-Human Handoff Model

Path A: TypeScript SDK (agentlaunch-sdk)

Install

Configure

Create your first token

Get a handoff link after deployment

Path B: CLI (agentlaunch)

Install

Configure

Create your first token

Create and deploy a new agent

Path C: MCP Server (agent-launch-mcp)

Install

Configure in Claude Code

Configure in Cursor

Create your first token via MCP

Path D: Autonomous On-Chain Trading

Requirements

Setup

Quick example (SDK)

Quick example (CLI)

Quick example (MCP)

Next Steps

Path A: TypeScript SDK (`agentlaunch-sdk`)

Path B: CLI (`agentlaunch`)

Path C: MCP Server (`agent-launch-mcp`)