OpenHands

What it does

OpenHands takes in engineering signals such as GitHub pull requests, Slack mentions, Jira or Linear issues, CI failures, and schedules, then runs agents that plan and carry out tasks across a codebase. Outputs include code changes, draft pull requests, review comments, triage notes, and remediation summaries, with engineers keeping control over review and merge.

Teams use it to automate repetitive software lifecycle work, such as triaging new issues, reviewing pull requests, fixing failed CI checks, and keeping documentation current. A distinguishing feature is model and agent flexibility: it can run its own agent or others such as Claude Code, Codex, and Gemini CLI, and it supports both closed and open-weight models, locally, in the cloud, or inside a VPC. Agent Canvas is the open-source browser client, and the SDK and Agent Server expose execution through APIs.

It suits engineering and platform teams that want an inspectable, self-hostable agent platform. Agent Canvas is free and open source under MIT, and self-hosting requires your own LLM key. The Teams plan is listed at $500 per month for up to 10 users, with additional seats at $50 per month each, and enterprise pricing is custom. The Teams plan is capped at 5,000 sessions per month, and model access must be supplied by you.

What it can do

• Trigger agent workflows from GitHub pull requests, Slack mentions, Jira or Linear issues, CI failures, and schedules
• Review pull requests and post findings as inline or summary comments
• Triage new issues, suggest labels, and find duplicates
• Investigate CI failures and open draft pull requests with proposed fixes
• Update documentation and codebase context as code changes
• Run agents in parallel and connect internal tools through MCP

Works with

• GitHub
• Slack
• Jira
• Linear
• MCP

Example workflows

Triage new Linear issues

Input: a newly created Linear issue. Steps: the agent classifies it, suggests labels, checks for duplicates, and asks clarifying questions. Output: a triaged issue with suggested labels and questions for the reporter.

Review labeled pull requests

Input: a pull request receiving a configured label. Steps: the agent inspects the full PR context and posts an AI review comment once per label event. Output: a review comment on the pull request for engineers to assess.

Before you choose

• Model access and compute costs depend on your own LLM key and infrastructure, not on the OpenHands subscription.
• The Teams plan is capped at 5,000 sessions per month; Enterprise usage is uncapped.
• Engineers remain responsible for reviewing and merging agent-generated changes.

Listed under

• Coding
• Code review and testing
• Coding agents
• DevOps and observability
• Automation
• Ci/cd
• Code review
• Coding agent
• Developer tools
• Open source

Get started

Agent Canvas install instructions are linked from the pricing page; the GitHub repository is the open-source entry point.

1. Download Agent Canvas or install the open-source project from GitHub.
2. Connect Agent Canvas to a local, self-hosted, Cloud, or Enterprise backend.
3. Add your own LLM key and select the agent or model to run.

Teams includes up to 10 users at $500/month, with additional seats at $50/month each. The 30-day trial applies to the Teams plan. Enterprise is custom. Model API and compute costs are separate.

30-day free trial.

Source: openhands.dev/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• OpenHands | Open Source AI Coding Agent Platform ↗www.openhands.dev
• OpenHands Pricing ↗www.openhands.dev
• Introduction - OpenHands Docs ↗docs.openhands.dev

Company website

openhands.dev: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.