How to Get Started with OpenHands: Local, Cloud, and Enterprise Options

OpenHands is an open-source AI coding agent platform you can start locally, run through OpenHands Cloud, or scale with enterprise controls. The right starting point depends on how much setup you want to own and whether your team needs governance. If you just want to try an agent against a repository, start local or with the cloud trial path. If you need scheduled automations across GitHub, Slack, Jira, or Linear with team-level controls, look at the cloud and enterprise options and check the pricing page for current plans.

What OpenHands actually is

OpenHands is a platform for building and running AI coding agents. Its own framing is "the open platform for agentic software development," with the tagline "Build it once. Let agents keep it running."

The core idea: you define an agent or automation once, and it keeps executing engineering work on a trigger — a new pull request, a Slack mention, a scheduled time — rather than you manually kicking off every task.

Key characteristics from the site:

• Open source — you can run it yourself.
• Bring your own agent and models — the platform is not locked to a single model provider ("Your agents. Your models. Your runtime.").
• Automation-first — the product centers on putting "the repetitive parts of the SDLC on autopilot."
• Integrations — triggers and actions connect to GitHub, Slack, Jira, Linear, schedules, and other supported integrations.

Your three starting paths

The site presents "Get OpenHands," "Try it now," and "OpenHands Cloud" as entry points, plus a "Talk to us" route that fits the enterprise conversation.

Path 1: Start locally

Choose this if you want to evaluate the agent, keep code and runtime on your own infrastructure, or use your own models without a hosted dependency.

The local entry screen is a build prompt — "Let's Start Building! What do you want to build?" — with a model selector and workspace/plugins options. In practice:

1. Point OpenHands at a repository or start a new project.
2. Select the model you want to use.
3. Give it a task in natural language and let it work in the workspace.

This is the lowest-commitment way to see whether the agent fits how you work before wiring up automations.

Path 2: OpenHands Cloud

Choose this if you want automations running without managing the runtime yourself. The cloud surface is where the automation features are most visible.

From the interface, you can:

• Start from a template or build your own automation.
• Trigger workflows from GitHub, Slack, Jira, Linear, schedules, and supported integrations.
• Watch recent automation activity — the dashboard shows runs like "PR Triage Digest — Weekdays at 09:00" and "Nightly Security Pass — Daily at 01:30" with status (Successful, Running).

A useful detail: automations carry a name, a schedule or trigger, a target repository, and a status, so you can see at a glance what has run and what is in flight.

Path 3: Enterprise controls

Choose this if you need governance and the ability to scale agent usage across a team or organization. The site explicitly lists "enterprise controls" as a capability, and offers a "Talk to us" contact path. If your blockers are access control, auditability, or rolling agents out beyond a pilot, that is the conversation to have rather than trying to bolt governance onto a local setup.

What you can automate once you're set up

The platform ships recommended automations and lets you customize your own. Concrete examples shown on the site:

• Slack channel monitor — watch a channel for mentions and open a conversation with the message context.
• Slack standup digest — collect standup replies and post a weekday digest to the team channel.
• PR Review on Open — review newly opened pull requests and post findings as inline comments.
• PR Triage Digest — review new PRs in a repo, identify risky changes, summarize likely impact, and produce a priority-ordered digest for the review channel (example: weekdays at 09:00).
• Nightly Security Pass — scan a repository for known vulnerabilities, outdated dependencies, and insecure code patterns, and produce a prioritized remediation summary (example: daily at 01:30).
• Docs Sync on Push — runs on every push.
• Linear issue triage — classify new issues, suggest labels, find duplicates, ask clarifying questions.
• Jira incident triage — watch for new bugs/incidents, gather logs and recent changes, suggest severity, route to the right owner.

The pattern across all of them: an event or schedule triggers the agent, the agent does analysis or a change, and the result lands where the team already works (a PR comment, a Slack channel, an issue).

How to choose

• Just evaluating? Start local. Lowest setup, and you can test your own model.
• Want automations without running infrastructure? OpenHands Cloud, starting from a template.
• Need governance across a team? Enterprise controls, via the "Talk to us" path.
• Cost-sensitive? There is a Pricing page linked from the site — check it for current plan details rather than assuming any tier is free.

Common sticking points

• Trigger setup is where most of the work is. The agent logic is reusable, but each automation needs its integration (GitHub, Slack, Jira, Linear) connected and scoped correctly before it fires.
• Model choice affects behavior. Since you bring your own model, results on the same task can differ by model — pick one deliberately rather than defaulting.
• Start with one automation, not ten. The templates make it tempting to enable everything at once; a single high-value automation (PR review or nightly security) is easier to judge and tune.
• "Open source" and "free" are not the same claim. The platform is open source; pricing for hosted or enterprise offerings is a separate question — see the Pricing page.