Skip to content

tonyoconnell/agent-launch-toolkit

Repository files navigation

AgentLaunch Toolkit

Build, deploy, and tokenize AI agents in one conversation.

  git clone agent-launch-toolkit && cd agent-launch-toolkit
  npm install && cp .env.example .env
  # Add your Agentverse API key
  claude

  > /build-agent
  > "Build a price monitoring agent and tokenize it as $PMON"

  Done. Agent running on Agentverse. Token live on-chain. Handoff link ready.

Full Tutorial: Launch Your First Token in 10 Minutes


Features

For Claude Code Users

Open this repo in Claude Code and everything works out of the box:

  • 5 slash commands/build-agent, /deploy, /tokenize, /market, /status
  • 13+ MCP tools — list tokens, calculate prices, deploy agents, create tokens, post comments
  • 3 auto-loaded rules — platform constants, Agentverse API patterns, uAgent Chat Protocol
  • Pre-configured MCP server.claude/settings.json already wired up

For CLI Users

npx agentlaunch create               # Full interactive wizard
npx agentlaunch scaffold MyBot       # Generate from template
npx agentlaunch deploy agent.py      # Deploy to Agentverse
npx agentlaunch tokenize agent1q...  # Create token + handoff link
npx agentlaunch list --sort trending # Browse tokens
npx agentlaunch status 0x...         # Check price and progress
npx agentlaunch holders 0x...        # Holder distribution
npx agentlaunch comments 0x...       # Read/post comments
npx agentlaunch config show          # Current environment + URLs
npx agent-launch init                # Install toolkit into any project

For SDK Users (TypeScript)

import { AgentLaunch } from 'agentlaunch-sdk';

const al = new AgentLaunch();

// Tokenize an agent
const token = await al.tokens.tokenize({
  name: 'PriceBot', symbol: 'PBOT',
  description: 'Monitors FET price',
  agentAddress: 'agent1q...',
  chainId: 97,
});

// Market data
const price = await al.market.getPrice('0x...');
const buy = await al.market.calculateBuy('0x...', '100');
const holders = await al.market.getHolders('0x...');

// Generate links for humans
const deployLink = al.handoff.generateDeployLink(42);
const buyLink = al.handoff.generateBuyLink('0x...', 100);
const sellLink = al.handoff.generateSellLink('0x...', 500);

For Any Language (Headless API)

API_URL=https://launchpad-backend-dev-1056182620041.us-central1.run.app

curl -X POST $API_URL/agents/tokenize \
  -H "Content-Type: application/json" \
  -H "X-API-Key: av-xxx" \
  -d '{"name":"MyBot","symbol":"MBOT","chainId":97}'

# List tokens
curl $API_URL/agents/tokens

# Token details
curl $API_URL/agents/token/0x...

Quick Start (2 Minutes)

Option A: Clone This Repo

git clone https://github.com/tonyoconnell/agent-launch-toolkit.git
cd agent-launch-toolkit
npm install
cp .env.example .env
# Paste your Agentverse API key in .env
claude

Then say: /build-agent

Option B: Add to Existing Project

cd my-project
npx agent-launch init
cp .env.example .env
# Paste your Agentverse API key in .env
claude

The init command installs .claude/ rules, skills, MCP config, .env.example, and Cursor rules. It merges into existing config files without overwriting.


What's Inside

Package npm Description
SDK agentlaunch-sdk TypeScript client for all API operations
CLI agentlaunch-cli 10 commands, full lifecycle, JSON output mode
MCP Server agent-launch-mcp 13+ tools for Claude Code and Cursor
Templates agentlaunch-templates 6 production-ready agent code generators

Agent Templates

Template Description
custom Blank Chat Protocol boilerplate
price-monitor Watch token prices, send alerts on thresholds
trading-bot Buy/sell signal generation
data-analyzer On-chain data analysis pipelines
research Deep dives, reports, knowledge retrieval
gifter Treasury wallet with FET reward distribution

The Lifecycle

  1. SCAFFOLD             2. DEPLOY              3. TOKENIZE            4. TRADE
  ──────────             ─────────              ───────────            ─────────
  Pick a template    ->  Upload to          ->  Create token       ->  Human opens
  Generate agent.py      Agentverse             record on              handoff link,
  Customize logic        Start agent            AgentLaunch            connects wallet,
                         Get agent1q...         Get handoff link       signs deploy tx
                                                                       Token is LIVE

Slash Commands (Claude Code)

Command What It Does
/build-agent Full guided flow: ask requirements, pick template, scaffold, deploy, tokenize
/deploy Deploy an agent.py to Agentverse hosting
/tokenize Create a token record for an existing agent, return handoff link
/market Browse tokens, check prices, see trending, calculate buy/sell
/status Check token price, progress, holder count, deployment status

MCP Tools (Claude Code / Cursor)

Pre-configured in .claude/settings.json. Available as soon as you open Claude Code.

Discovery

Tool Description
list_tokens Browse tokens with status/chain/category filters
get_token Full details for a token by address or ID
get_platform_stats Total volume, token count, trending

Price Calculations

Tool Description
calculate_buy Preview buy: tokens received, fee, price impact
calculate_sell Preview sell: FET received, fee, price impact

Token Operations

Tool Description
create_token_record Create pending token, get handoff link
create_and_tokenize Full lifecycle: scaffold + deploy + tokenize
update_token_metadata Update description, logo, links

Handoff Links

Tool Description
get_deploy_instructions Structured deploy instructions for humans
get_trade_link Pre-filled buy/sell URL for humans

Agentverse

Tool Description
deploy_to_agentverse Deploy Python agent, upload code, start
check_agent_logs Read agent execution logs
stop_agent Stop a running agent

Code Generation

Tool Description
scaffold_agent Generate agent project from template
get_agent_templates List available templates

Social

Tool Description
get_comments Read comments on a token
post_comment Post a comment on a token

CLI Commands

Command Description
agentlaunch create Interactive wizard: scaffold, deploy, tokenize in one flow
agentlaunch scaffold <name> Generate agent project from template
agentlaunch deploy [file] Deploy agent.py to Agentverse
agentlaunch tokenize <agent> Create token record + handoff link
agentlaunch list List tokens (filter, sort, paginate)
agentlaunch status <address> Token price, progress, holders
agentlaunch holders <address> Holder distribution
agentlaunch comments <address> Read or post comments
agentlaunch config show Show environment, URLs, chain, API key
agentlaunch config set-key <key> Store API key
agentlaunch init Install toolkit into any project

All commands support --json for machine-readable output.


SDK API

import { AgentLaunch } from 'agentlaunch-sdk';
const al = new AgentLaunch({ apiKey: 'av-...' });

Namespaces

Namespace Methods
al.tokens tokenize(), get(), list()
al.market getPrice(), getHolders(), calculateBuy(), calculateSell(), getStats()
al.handoff generateDeployLink(), generateTradeLink(), generateBuyLink(), generateSellLink()
al.agents authenticate(), getMyAgents(), importFromAgentverse()

Standalone Functions

Every method is also available as a standalone function:

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

Environment Configuration

Copy .env.example to .env:

AGENTVERSE_API_KEY=av-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

# Environment: dev (default) or production
AGENT_LAUNCH_ENV=dev

Live URLs

Environment Frontend Backend API
Dev (default) https://launchpad-frontend-dev-1056182620041.us-central1.run.app https://launchpad-backend-dev-1056182620041.us-central1.run.app
Production https://agent-launch.ai https://agent-launch.ai/api

The toolkit defaults to dev URLs — no configuration needed. To switch to production, set AGENT_LAUNCH_ENV=production in your .env. You can also override URLs directly:

AGENT_LAUNCH_API_URL=https://launchpad-backend-dev-1056182620041.us-central1.run.app
AGENT_LAUNCH_FRONTEND_URL=https://launchpad-frontend-dev-1056182620041.us-central1.run.app

URL resolution priority:

  1. AGENT_LAUNCH_API_URL / AGENT_LAUNCH_FRONTEND_URL (direct override)
  2. AGENT_LAUNCH_ENV=production (uses agent-launch.ai)
  3. Default: dev Cloud Run URLs (shown above)

Platform Details

Property Value
Deploy Fee 120 FET (read dynamically, can change via governance)
Graduation 30,000 FET raised triggers auto DEX listing
Trading Fee 2% on every buy/sell, 100% to protocol treasury
Token Supply 800,000,000 per token
Default Chain BSC (Testnet: 97, Mainnet: 56)
Auth X-API-Key header with Agentverse API key
Dev Frontend https://launchpad-frontend-dev-1056182620041.us-central1.run.app
Dev Backend https://launchpad-backend-dev-1056182620041.us-central1.run.app

The 2% trading fee goes 100% to the protocol treasury (REVENUE_ACCOUNT). There is no creator fee split.


Architecture

  .env (AGENTVERSE_API_KEY)
    |
    +--- SDK (TypeScript HTTP client)
    |      |
    |      +-- CLI (interactive + scripted commands)
    |      +-- MCP (13+ tools for Claude Code / Cursor)
    |      +-- Templates (6 agent code generators)
    |
    +--- Claude Code
           reads CLAUDE.md
           loads .claude/rules + skills
           has MCP tools auto-configured
           |
           /build-agent
             1. Scaffold (templates)
             2. Deploy (Agentverse API)
             3. Tokenize (AgentLaunch API)
             4. Handoff link -> Human signs -> Token LIVE

See docs/architecture.md for package dependency diagrams.


Documentation

Doc Description
Tutorial Step-by-step: launch your first token in 10 minutes
SDK Reference All TypeScript methods and types
CLI Reference All commands, flags, and examples
MCP Tools All tools with input/output schemas
Getting Started Three paths: SDK, CLI, MCP
Architecture Package diagrams and data flow
Agent Integration Building agents that launch tokens

Development

npm install        # Install all workspace dependencies
npm run build      # Build all 4 packages
npm run test       # Run all tests (78 tests across SDK + CLI)
npm run clean      # Clean all dist/ directories

Links


License

MIT — see LICENSE.

About

AgentLaunch Toolkit — SDK, CLI, MCP server, and templates for building AI agent token launches

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors