Skip to content
1 change: 1 addition & 0 deletions docs.json
Original file line number Diff line number Diff line change
Expand Up @@ -199,6 +199,7 @@
"pages": [
"overview/community",
"overview/contributing",
"overview/strategy",
"overview/faqs",
"openhands/usage/troubleshooting/feedback"
]
Expand Down
193 changes: 193 additions & 0 deletions overview/strategy.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,193 @@
---
title: OpenHands Commercial Open Source Strategy
description: OpenHands' strategy to become the open standard for AI coding agents
---

## Vision


OpenHands' vision is to become **the open standard for AI coding agents** — a foundation for AI software development that is transparent, extensible, and community-driven.


## Problems We Solve


### "Building With" OpenHands: Use OpenHands as an Inner Loop local AI coding agent


**Problem:** Developers need a reliable and powerful AI coding agent to boost productivity.


- Designed for most **individual developers**
- **Ad-hoc usage** focused on local or small-team development
- Solves **developer productivity** challenges
- Provides **model-agnostic AI code assistance** that integrates locally
- Operates in a **highly competitive** landscape with multiple alternatives


### "Building On" OpenHands: Use OpenHands to solve repetitive, Outer Loop engineering tasks


**Problem:** Organizations need a way to reduce the toil and burden of recurring engineering tasks that distract teams from roadmap execution. They seek solutions that can scale to nearly every development team in the organization.

- Designed for **teams and organizations — these tend to be Commercial entities**
- Within organizations, used by both **Developers** who integrate AI into their coding process, and **Agent Builders** who build workflows on top of OpenHands
- Supports **repeatable, cloud-based workflows and automations.** Sometimes this is referred to as "Outer Loop" engineering tasks, such as PR reviews, security remediation, and documentation updates.

Check warning on line 35 in overview/strategy.mdx

View check run for this annotation

Mintlify / Mintlify Validation (allhandsai) - vale-spellcheck

overview/strategy.mdx#L35

Did you really mean 'automations'?
- Enables **API-driven integration** into applications and development pipelines
- Solves problems around **scalability, extensibility, and control**
- Focused on enabling **custom agent orchestration**


**Strategic Shift:**


We're evolving OpenHands to support "Building On" use cases — turning OpenHands into a platform for scaling 100s of agents that address repetitive engineering problems. Some examples of this:
- We refactored OpenHands into a fully open source, [Software Agent SDK](https://github.com/OpenHands/software-agent-sdk/)
- We launched a [Plugin Spec](https://docs.openhands.dev/sdk/guides/plugins) to enable packaging of Agent Skills, Sub-agents, Hooks, MCP servers, and more to help 'bundle' use cases.
- We launched [Use Case documentation](https://docs.openhands.dev/openhands/usage/use-cases/overview) and an [Extensions repo](https://github.com/OpenHands/extensions/tree/main) with Agent Skills, plugins, and SDK examples

## Example Use Cases — Repetitive Engineering Tasks


OpenHands is used across a wide range of powerful, scalable use-cases in the SDLC process:


- **Maintenance**
- Example: 30x throughput on CVE resolution
- Example: Automatic documentation and release notes
- **Modernization**
- Example: Adding type annotations to an entire Python codebase
- Example: Refactoring a monolithic Java application to microservices
- **Migration**
- Example: Upgrading 1000s of jobs from Spark 2 to Spark 3
- Example: Moving from Redux to Zustand

Check warning on line 63 in overview/strategy.mdx

View check run for this annotation

Mintlify / Mintlify Validation (allhandsai) - vale-spellcheck

overview/strategy.mdx#L63

Did you really mean 'Zustand'?
- **Tech Debt**
- Example: Detecting and deleting unused code
- Example: Adding error handling based on production logs
- **Automated Testing, Bug Fixing, Documentation**, etc.


## Target Customers for OSS and Commercial


### Open Source (OSS)


- **Goal:** Empower individual developers with a model-agnostic, open, and extensible coding agent that integrates seamlessly into their workflows.
- **Primary audience:** Individual developers on the bleeding edge of software development; AI hobbyists; "Agent Builders" who want to use the Software Agent SDK to build their own coding agent workflows.
- **Use cases:** Personal productivity, local development, experimentation, and contributing to open AI infrastructure.
- **Relevant Projects/Products:**
- OpenHands Software Agent SDK
- OpenHands Local GUI
- OpenHands CLI
- OpenHands Cloud (app.all-hands.dev)

### Commercial

- **Primary audience:** Mid-to-large enterprises and technical teams.
- **Goal:** Help organizations safely leverage AI at scale to modernize, refactor, and maintain complex systems across 100s of development teams.
- **User personas:**
- **Developers w/ AI experience:** Developers who already may be familiar with pairing with AI to accomplish work, but ready to begin delegating wholesale tasks to AI agents for maximum productivity.
- **Agent Builders:** Developers on the edge of AI engineering who are looking to integrate agents into workflows or applications.
- **Buyer personas:** AI / Platform Engineering Leaders
- **Use cases:**
- "An OpenHands agent for every developer": Multi-user collaboration, scalability, governance, and security
- Refactoring and code modernization
- Embedding AI agents into apps and workflows
- **Relevant Projects/Products:**
- OpenHands Software Agent SDK
- OpenHands CLI
- OpenHands Enterprise


## Dividing Lines Between OSS and Commercial

_OpenHands' CEO and founder discusses the lines between OSS and Commercial in this blog post: [Walking the Line with Commercial Open Source (June 2025)](https://openhands.dev/blog/walking-the-line-with-commercial-open-source)_


OpenHands follows an **Open Core** model:


- **Open Source Core:**

The foundation of OpenHands is open and free for all developers. Most new features default to open source unless they explicitly solve enterprise-specific needs.

- **Enterprise Offering:**

Focused on mid-to-large enterprises, addressing needs around multi-user management, scalability, and security.



Our investment is roughly **50/50** between open source and commercial development.


### **Feature Selection Principle**


- **Developer problems → Open source by default:**

Features that solve an individual developer's needs belong in the open source core (e.g., productivity, model integration, developer tooling).

- **Buyer problems → Commercial features:**

Features designed to solve organizational or managerial challenges are commercial (e.g., organizational management, collaboration, scalability).



**Example:**


- Commercial feature: Organizational Management — features for managing OpenHands access across hundreds of developers
- Open Source: MCP support, Secrets Support, Planning Agent mode, etc.



## Unique Differentiators


- **Open Source Foundation and Community:** Transparent, extensible, and a community-driven approach to maintaining OpenHands.
- **Model-Agnostic Architecture:** Works with any LLM — the durable value lies in the agent layer, not the model.
- **Cloud Agents:** Run OpenHands in sandboxed containers on remote agent servers -- whether locally or on cloud infrastructure.

Check warning on line 150 in overview/strategy.mdx

View check run for this annotation

Mintlify / Mintlify Validation (allhandsai) - vale-spellcheck

overview/strategy.mdx#L150

Did you really mean 'sandboxed'?
- **Focus on Repetitive Tasks:** Automates ongoing engineering work — refactoring, maintenance, dependency upgrades — where AI delivers compounding value.
- **Dual Persona Alignment:** Serves both individual developers ("Build With") and Agent Builders ("Build On") through a unified ecosystem.


## Strategic Roadmap Pillars


1. **Bet on "Cloud Agents" as the best way to work with AI coding agents**

Prioritize sandboxed runtimes for scalable, persistent, autonomous systems. When run on a remote agent server, it may be referred to as a "Cloud Agent".

Check warning on line 160 in overview/strategy.mdx

View check run for this annotation

Mintlify / Mintlify Validation (allhandsai) - vale-spellcheck

overview/strategy.mdx#L160

Did you really mean 'sandboxed'?

Check warning on line 160 in overview/strategy.mdx

View check run for this annotation

Mintlify / Mintlify Validation (allhandsai) - vale-spellcheck

overview/strategy.mdx#L160

Did you really mean 'runtimes'?

2. **Shifting towards "Build On" to optimize for repetitive, high-leverage work**

Emphasize ongoing tasks like maintenance and refactoring over one-off code generation.

3. **Maintain a high performance, model-agnostic approach**

Continue to integrate multiple models and support extensibility as the LLM ecosystem evolves.

4. **Bridge both developer needs and enterprise needs**

Build pathways that help individual developers grow into Agent Builders for their organization — from "Build With" mode to "Build On" mode.

5. **Foster a thriving Open Source community**

Grow contributor engagement, maintain transparent governance, and accelerate open development with clear & specific roadmaps.

Check warning on line 176 in overview/strategy.mdx

View check run for this annotation

Mintlify / Mintlify Validation (allhandsai) - vale-spellcheck

overview/strategy.mdx#L176

Did you really mean 'roadmaps'?



## Commitment to Open Source


We believe open source works best with clear stewardship and active participation.


- We maintain the **roadmap** and **good first issues** to guide community contributions.
- We **review pull requests diligently** and recognize valuable contributions.
- We maintain **transparent metrics** through our public [OpenHands PR Review dashboard](https://prdashboard.openhands.dev/)

### Our long-term commitment:


To make OpenHands the **best open source generalized AI coding agent**, serving as the backbone of transparent, collaborative AI development.
Loading