OpenHands

What it does

Writes code

YesTurns a scoped issue into a reviewable pull request, tests included.Turn well-scoped issues into tested, reviewable pull requests using your existing tools.openhands.dev/services ↗

Debugs and fixes

YesInvestigates failing builds, tests and production errors and can implement the fix, not just describe it.Investigate failed builds and tests, identify likely causes, and propose or implement fixes.openhands.dev/services ↗

Writes tests

YesTest generation and coverage expansion is a named workflow; a separate QA agent also exercises the running software rather than the test suite.Generate tests, expand coverage, and keep test suites aligned as the codebase changes.openhands.dev/services ↗

Reviews code

YesA GitHub Actions PR-review agent posts inline comments with priority labels on someone else's diff.OpenHands provides a GitHub Actions workflow powered by the Software Agent SDK that automatically reviews pull requests and posts inline comments…docs.openhands.dev/openhands/usage/use-… ↗

Whole-codebase work

PartlyRepository-wide migrations are a documented use case, but decomposing systems spanning hundreds of files relies on the Large Codebase SDK, which the plan table lists only under Enterprise.the OpenHands large codebase SDK, which combines agentic intelligence with static analysis tools to decompose large projects and orchestrate parallel…docs.openhands.dev/openhands/usage/use-… ↗

Ships it

PartlyThe documented end state of an agent run is a pull request left for human review; no deploy-to-production step is claimed.When something happens, the right agent fires, does the work, and opens a pull request.openhands.dev/enterprise ↗

Done-for-you service

PartlyYou run the software, but All Hands AI also sells advisory and implementation engagements that design and deploy the agents alongside your team.We work alongside your team to design and deploy production-ready agents that are integrated with your environment and workflows, on an OpenHands…openhands.dev/services ↗

How much it does unattended

Runs autonomously

YesEvent-triggered automations run start to finish unattended; headless CLI runs are forced into always-approve mode.From event to PR, no human in the loop.openhands.dev/enterprise ↗

Multi-agent

YesA parent agent delegates to registered sub-agent types through the Task tool set; sub-agents run synchronously and can be resumed by task id.The TaskToolSet lets a parent agent launch sub-agents that handle complex, multi-step tasks autonomously.docs.openhands.dev/sdk/guides/task-tool… ↗

Agent permissions

YesThree confirmation policies: approve everything, nothing, or only actions an LLM security analyzer rates risky. The same modes are switchable mid-session in the CLI and IDE surfaces.Confirmation policy controls whether actions require user approval before execution.docs.openhands.dev/sdk/guides/security ↗

Which models it runs on

Claude

YesClaude Sonnet, Opus and Haiku are sold through the OpenHands LLM provider at listed per-million-token rates, and are also reachable with your own Anthropic key.claude-sonnet-4-5-20250929 $3.00 $0.30 $15.00 200,000 64,000docs.openhands.dev/openhands/usage/llms… ↗

GPT

YesGPT-5, GPT-5 Codex and GPT-5 mini are sold through the OpenHands LLM provider at listed per-million-token rates, alongside o3 and o4-mini.gpt-5-2025-08-07 $1.25 $0.125 $10.00 272,000 128,000docs.openhands.dev/openhands/usage/llms… ↗

Open models

YesOpen-weight models are recommended and benchmarked by the vendor, with documented setup for serving them locally through LM Studio, Ollama, vLLM and SGLang.These open or open-weight models have good OpenHands Index scores or are recommended for local OpenHands setups:docs.openhands.dev/openhands/usage/llms… ↗

Model choice

YesYou pick the provider and model; routing goes through LiteLLM, with vendor guides for Bedrock, Azure, Google, Groq, OpenRouter, Moonshot and local servers.For a full list of the providers and models available, please consult the litellm documentation.docs.openhands.dev/openhands/usage/llms… ↗

Where you use it

In your editor

PartlyZed, JetBrains and Toad connect natively over the Agent Client Protocol and VS Code via a community extension, but the vendor labels the whole integration experimental.IDE integration via ACP is experimental and may have limitations.docs.openhands.dev/openhands/usage/cli/… ↗

On the command line

YesA terminal CLI is listed on every plan and also runs headlessly for scripting.Run, inspect, and automate OpenHands agents directly from the terminal or headlessly in your pipelines with full visibility and control.openhands.dev/product ↗

In your pipeline

YesShips GitHub Actions plugins for PR review and QA that fire on pull-request events, plus a headless CLI mode with JSONL output for pipelines.Triggers automatically when PRs are opened or when you request a reviewdocs.openhands.dev/openhands/usage/use-… ↗

In a browser

YesA web GUI (Agent Canvas) is listed on every plan, locally and hosted.Plan, run, and review agent work together in a collaborative UI.openhands.dev/product ↗

Whose machine it runs on

Self-hosted

YesRuns locally for free, and Enterprise installs into your own Kubernetes cluster or VPC with no shared tenancy.Deploy on your own Kubernetes infrastructure. Your code never leaves your environment. No shared tenancy.openhands.dev/enterprise ↗

Open source

YesThe open-source distribution is MIT licensed; the docs warn that each repository in the ecosystem carries its own licence, and Enterprise is source-available rather than open.OpenHands Open Source is a full, MIT licensed version of OpenHands that is designed for a single user and can run locally on your machine with your…openhands.dev/pricing ↗

What it costs to run

How it meters

YesMetered per token at the upstream provider's own rates with no markup, paid as you go; there is no per-seat subscription.which provides multiple model options at cost, with no markup on a pay-as-you-go basis.openhands.dev/pricing ↗

Free tier

YesTwo free routes with no call: the local open-source build, and a free hosted Individual plan capped at 10 conversations a day.Free, local OpenHands for getting startedopenhands.dev/pricing ↗

API access

YesA REST API starts and polls conversations, and the plan table ticks Cloud APIs on all three tiers.OpenHands Cloud provides a REST API that allows you to programmatically interact with OpenHands.docs.openhands.dev/openhands/usage/clou… ↗

MCP server

YesAgent Server exposes an OpenAI-compatible chat-completions endpoint that another agent or client can call, holding conversation state behind a returned header.The agent-server exposes an OpenAI-compatible /v1/chat/completions endpoint so clients that already speak the OpenAI protocol can call an OpenHands…docs.openhands.dev/sdk/guides/agent-ser… ↗

Bring your own key

YesYour own LLM keys work on every tier; Enterprise additionally runs on your own compute in your own VPC.Yes! OpenHands Open Source and OpenHands Cloud plans offer support to bring your own LLM keys.openhands.dev/pricing ↗

Buying it for a team

A company can buy it

YesOrganizations are sold to companies as a commercial subscription, and Enterprise is a contracted purchase quoted by sales.Organizations is a commercial feature available with an OpenHands Cloud subscription or OpenHands Enterprise.docs.openhands.dev/openhands/usage/clou… ↗

Seat model

NoEnterprise is explicitly not licensed per seat and carries unlimited users; the two free tiers are each capped at one user.Use your own compute resources and LLM API keys. No per-seat licensing.docs.openhands.dev/enterprise ↗

Pooled budget

PartlyCredits are pooled across all members, but only inside an Organization, which requires a paid Cloud subscription or Enterprise.Shared Credits - Pool OpenHands Cloud credits across all organization members.docs.openhands.dev/openhands/usage/clou… ↗

Admin controls

PartlyAdmins can constrain the secrets, tools and domains an agent may reach and halt risky actions, but that control plane is Enterprise; Cloud Organizations only reach as far as default LLMs and budget caps.Define exactly which secrets, tools, and domains each agent can reach. Detect prompt injection. Halt risky actions for human review.openhands.dev/enterprise ↗

Audit log

PartlyPer-action audit logs and run attribution are part of the Enterprise control plane; nothing equivalent is claimed for the free or Individual tiers.Complete audit logs for every agent actionopenhands.dev/enterprise ↗

Single sign-on

PartlySAML/SSO ships only with the self-hosted Enterprise deployment; the vendor's own comparison table shows SAML absent from both Open Source and hosted Cloud.Identity & Access: Enterprise SAML/SSO for centralized authenticationdocs.openhands.dev/enterprise ↗

What happens to your code

Opt out of training

PartlySelf-hosting is the published route to keeping code out of vendor hands; no opt-out control is documented for the hosted service, whose privacy policy reserves the right to train on submitted content.All code and conversations stay on your infrastructure. Nothing leaves your environment.docs.openhands.dev/enterprise ↗

Data residency

PartlyEnterprise lets you place code and data on-premises, in your private cloud or in your VPC; for the hosted service the vendor names no region, country or jurisdiction, only 'OpenHands Cloud'.Deploy OpenHands on your own infrastructure—whether on-premises, in your private cloud, or in your VPC. You maintain full control over where your…docs.openhands.dev/enterprise ↗

Getting out

PartlyAccount content can be managed, deleted in-product, or deleted on written request; no data-export mechanism and no credit-expiry term is published, and payments are stated to be non-refundable.If you require your data to be deleted, please email contact@all-hands.dev.openhands.dev/privacy ↗

Not established: Certifications