# Alinery

> A new way to build. A Playbook gives you a ready-to-go team of agents you can steer as you discover what your work needs.

Alinery is a desktop app: A Playbook IDE for creating, inspecting, and steering teams of AI agents. Open source under Apache 2.0, for macOS on Apple Silicon and Linux on x86_64.

Playbooks define the process agents follow, their responsibilities, and how they hand work to one another. Agents can run dynamically as a task develops. Tasks hold shared context; sub-tasks fork that context for focused exploration that can be discarded or merged back.

Alinery includes OMP for agent sessions and a Terminal option for shell work. Models can be chosen per Playbook step. External agents can manage tasks, check progress, and inspect artifacts through MCP (Model Context Protocol). Markdown artifacts hold plans, decisions, and outputs for review. Tasks, sessions, and artifacts live in the repository's `.alinery/` directory.

Use the linked pages for detailed explanations and current product information. The talk is a video; the whitepaper is a PDF hosted in the public source repository.

## Start here

- [Homepage](https://alinery.ai/): Product overview and interactive examples of Playbooks, tasks, and sub-tasks.
- [Learn](https://alinery.ai/learn): Explanations of the core concepts and how they fit together.
- [Install Alinery](https://alinery.ai/#install): Installation instructions. Terminal command: `curl -fsSL https://cdn.alinery.ai/install.sh | bash`.
- [Installer source](https://cdn.alinery.ai/install.sh): Read the installation script.
- [Pricing](https://alinery.ai/pricing): The local app is $0. Founder is $7.99 with $5 in monthly AI credits and limited founder T-shirt incentives. Check this page for current plans.

## Learn the concepts

- [What is a Playbook?](https://alinery.ai/learn#what-is-a-playbook): Choose a process for a team of agents and steer their work as new information emerges.
- [Tasks](https://alinery.ai/learn#tasks): Organize work around a task's context, with multiple agents working within it.
- [Sub-tasks](https://alinery.ai/learn#sub-tasks): Fork task context to explore a problem, then discard the exploration or merge useful work back.
- [MCP](https://alinery.ai/learn#mcp): Connect external agents to Alinery's task and artifact tools.
- [Model and harness freedom](https://alinery.ai/learn#model-harness-freedom): Choose models for Playbook steps and learn about harness options, including OMP.
- [Playbook library](https://alinery.ai/learn#playbook-library): Start from an existing Playbook, adapt it, or create and share your own.

## Research and talk

- [Finite Agent Machines whitepaper (PDF on GitHub)](https://github.com/DeleganceAI/alinery/blob/main/whitepaper/finite-agent-machines.pdf): The paper on the repository's main branch; this URL follows updates to the PDF at the same path.
- [Talk: Playbooks vs Existing Agentic Approaches](https://alinery.ai/learn#talk): Embedded recording of Dustin's talk comparing Playbooks with other agentic approaches.

## Source and updates

- [Source repository and README](https://github.com/DeleganceAI/alinery#readme): Open-source implementation, setup, and project documentation.
- [Blog](https://alinery.ai/blog): Published articles about AI-assisted work and Alinery.
- [Blog RSS feed](https://alinery.ai/blog/rss.xml): Machine-readable feed of published articles.

## Optional

- [Privacy policy](https://alinery.ai/privacy): Data handling and privacy details.
- [Terms](https://alinery.ai/terms): Terms for using Alinery.
- [Sitemap](https://alinery.ai/sitemap.xml): Index of public pages and published blog posts.