Stacks 1.0: The freedom to choose your AI provider and model

Stacks 1.0 is here! Run multiple AI providers, use slash commands, save your AI chat history, and enjoy a totally overhauled experience for building with Stacks. Upgrade now and unlock the next level of AI-first development.

What is Stacks?

Stacks decentralized AI template-sharing platform built on Nostr, designed specifically to jumpstart development using templates and tools alongside LLMs. Anyone can publish a stack, which allows others to use their templates as a starting point for AI-first coding. MKstack, a tool for building Nostr apps with AI, is one such stack!

Stacks includes a CLI tool for managing and communicating with AI models through the terminal.

Featured stacks. Explore more on getstacks.dev

What’s New in Stacks 1.0?

Be your own open router: Easily configure and switch between multiple AI providers.
Quick upgrade: Just run stacks upgrade to get started (note: provider config will reset for a smoother setup experience).
Multiple providers at once: Use different providers for different models, or even your own custom models.
Per-project AI settings: Each project remembers its own AI settings—no more sharing models across users or projects.
Slash commands: Type / in any Stacks agent session to see all available commands. Change models, temperature, and more—mid-prompt!
AI chat history: All your AI sessions are saved in a new .ai folder (inspired by .git), so you can revisit and reload previous chats.
Quick model switch: Instantly swap between your 10 most recent models.
Overhauled menus: Every menu has been redesigned for a smoother, more intuitive experience.

Quickly swap between providers and models

Stacks 1.0 makes it easier than ever to manage multiple AI providers and models. Store your settings for providers like OpenRouter, PayPerQ, and more, and even use different providers for different models..

You can access models from any provider with a slash `/`. So, GPT-4o on OpenAI's official API would be called `openai/gpt-4o`, while accessing it on OpenRouter would be `openrouter/openai/gpt-4o`

Once you use a model, stacks will remember it and suggest it for easy swaps in the future.

Slash Commands: Power at Your Fingertips

Stacks 1.0 introduces a powerful new slash command system. Just type / in any chat with your Stacks agent to:

Switch models or providers instantly (/model, /providers)
List available tools (/tools)
Show or load chat history (/history, /load)
Adjust temperature (/temperature)
And more! Type /help or / to see all commands

Type / to access all commands in your Stacks agent session

AI Chat History: Never Lose Your Flow

Every chat with your AI agent is now saved in a .ai folder in your project—just like .git. Reload any previous session, pick up where you left off, or review your progress at any time. Each project keeps its own settings and history, so you can use GPT for one project and GLM for another, with no conflicts.

All your AI chat sessions are saved and reloadable in Stacks 1.0

How to Upgrade

Update your CLI: npm install -g @getstacks/stacks
In your project folder, run: stacks upgrade
Re-enter your provider config (it’s easier than ever!)
Start building: stacks agent

Upgrading to Stacks 1.0 is just one command away

A Foundation for the Future

The new .ai folder is inspired by .git and lays the groundwork for the next phase of AI-first, decentralized development. This is a major step toward our vision to #FreeGit and paves the way for Shakespeare Act 2 and beyond.

We’ve also moved provider config from agent.json to .mcp.json for a cleaner, more flexible setup.

Try It Now

Ready to experience the new Stacks? Upgrade today and unlock the most powerful, flexible AI-first development workflow yet.

Get Notified of New Releases

Try Stacks 1.0 →