OpenHands

Agent Canvas is the browser based control surface where developers manage conversations, files, terminals, and automations in one place. It supports not only the native OpenHands agent but also external agents like Claude Code, OpenAI Codex, and Gemini CLI through the Agent Client Protocol (ACP). This means you are not locked into a single agent. You can route tasks to the best agent for the job while keeping a single operational view, saving significant context switching time.

OpenHands turns signals from your engineering stack into end to end agent workflows. A new Jira ticket, a failed CI run, a Slack mention, or a Linear issue can automatically trigger an agent to investigate, write a fix, and open a pull request. These automations run in the background, even after you close your laptop, handling the repetitive parts of the software development lifecycle and freeing engineers to focus on higher value design and review work.

Unlike proprietary competitors that lock you into a single LLM provider, OpenHands connects to any model through LiteLLM. Teams can use Claude Opus for complex planning tasks, switch to a cheaper model for routine edits, and even run fully local models through Ollama or vLLM when sensitive code must never leave the network. This flexibility directly impacts both cost efficiency and data governance.

For organisations scaling agent use across multiple teams, the Enterprise tier provides a centralised governance layer. This includes role based access control (RBAC), SAML/SSO authentication, audit logs, usage monitoring, budget enforcement, and self hosted VPC deployment. Platform and security teams get full visibility into who is running agents, what they can access, where they execute, and what they spend.

The Large Codebase SDK maps dependencies across an entire system and orchestrates changes in the correct order. Multiple agents can safely work in parallel without conflicts, making OpenHands particularly effective for large enterprise codebases and legacy system modernisation projects that would overwhelm simpler tools.

OpenHands can watch for labelled pull requests, inspect the full PR context, and post an AI generated review comment. When CI pipelines fail, it can automatically detect the failure, inspect logs, identify the root cause, and open a fix PR. These features close the loop on two of the most time consuming bottlenecks in engineering workflows.