The Agent-Native Stack for Farcaster and Base

Build onchain mini apps in minutes with AI.

Lazer Mini Apps gives you a fast CLI and agent-native workflow for creating apps with platform adapters, web3 modules, and incremental feature activation built directly into your AI agent's toolkit.

Lazer CLI Farcaster + Base
$ bunx @lazer-tech/miniapp create my-app
$ cd my-app
$ bun dev
Works with Farcaster + Base
Use it from CLI, MCP, or skills

What this actually gives you.

Not another generic starter. This is opinionated tooling for mini apps: create the base app fast, give agents the right source references, and let humans or agents expand the project with less setup drift and less unused code.

Start with one command

Scaffold a clean mini app for Farcaster and Base without spending a day wiring adapters, manifests, or wallet plumbing. The base app is ready for humans and agents.

Build from references, not bloat

Generated apps include agent-readable source references for auth, web3, and more, so skills can implement only the code your app actually needs, tailored to your requests.

Built for humans and agents

Use the CLI directly, or open the project in your agent. Every app comes with built-in slash-commands like /auth and /farcaster so agents can build for you.

How it works.

Lazer works differently than a traditional template. Instead of starting with a bloated starter, it gives your agent the references it needs to build only the features you ask for.

01

Create

Scaffold a minimal app shell. It comes pre-configured with your chosen platforms and agent-readable source references for everything else.

02

Build

Tell your agent what to implement. It uses the built-in skills to surgically write auth, web3, or data logic directly into your source code.

03

Own

No black-box libraries or bloated templates. Your agent writes only the code you need, leaving you with a clean, editable, and auditable repo.

See the workflow.

Lazer transforms a basic shell into a feature-rich mini app through surgical, agent-driven implementations. Experience the workflow through this interactive simulation.

interactive workflow preview
LZ
Lazer Agent Harness v0.1agent mode · slash commands
Tip: type / to browse commands.
The demo is ready. Start with a slash command.
Type / to browse commands, then press Enter to run one.
branch mainmode react workflowstatus active
9:41
TokenPulse
TokenPulse

Track the chart. Make your pick.

ETH weeklySocial appBase + Farcaster
Weekly Prediction

Make the call before Friday close, then push your result back into the feed.

ETH
EthereumCommunity market
$4,182
Will ETH close above $4,200?Unlock a leaderboard spot.
Ends Fri
Contract missing

Onchain reads and writes are disabled. Run /contracts to deploy the prediction contract.

Built with Lazer.

See what's possible when you skip the boilerplate and move straight to building your core game logic and features.

Tile Riot
Base Sepolia · Coming Soon

Tile Riot

A fast-paced, fully onchain version of 2048 built on Base. Lazer provided the onchain foundations, allowing for a fast build of the game logic and smart contracts.

Visit site
Idle Perps
Base Sepolia · Coming Soon

Idle Perps

An idle trading game using real-time price feeds from Codex. Built using Lazer's data and auth skills to handle the heavy lifting.

Visit site

FAQ

Short answers for the questions a landing page should settle quickly.

What is Lazer Mini Apps?

It is an agent-native toolkit for building mini apps for Farcaster and Base. It scaffolds the app, ships skill references for major integrations, and helps agents implement the parts your app actually needs.

Why not just start from a generic Next.js starter?

Because mini apps need more than a blank app shell. You need platform adapters, manifests, wallet flows, design tokens, and a way to keep agent workflows consistent across different projects.

Who is this for?

It is for teams and developers building for Farcaster or Base who want to move faster by using AI agents for the heavy lifting of platform wiring and feature implementation.

Do I need to use AI agents to get value from it?

Mostly, yes. Lazer is designed to be an agent's favorite toolkit. While the CLI gives you the base app, the real leverage comes from agents using the shipped skills to implement features quickly and correctly.

Can AI agents use this directly?

Yes. The CLI supports MCP, machine-readable manifests, and installable skill files so agents like Claude, Cursor, and Gemini can discover and call commands through typed interfaces.

What kinds of features can it help me add?

Auth, EVM tooling, Solana tooling, token swaps, Codex data, Postgres, Farcaster integrations, Base integrations, and smart contracts via Foundry or Anchor.