OpenHands

AI security improves when the infrastructure around the model is as open as the model itself. Once agents can access repos, execute commands, use credentials, and call tools, security has to extend across identity, permissions, sandboxes, policies, logs, and execution. We’re excited to join the Open Secure AI Alliance alongside @NVIDIA and other industry leaders working together, in the open, to advance AI security across the whole stack. hubs.la/Q04sZGkv0

The same LLM can deliver very different results depending on the harness running it. Tools, context, execution, and verification all matter. That’s the premise behind OpenHands: model choice backed by the infrastructure to put agents to work. Check out the full article from iTech Guides: hubs.la/Q04DW_W30

AI coding agents are becoming shared infrastructure. Managing them requires more than a capable model. We’ve shipped 5 OpenHands Enterprise releases with better identity management, automation governance, spend controls, and Azure DevOps support. What’s new 👇 hubs.la/Q04DXrYY0

Join us tomorrow for the OpenHands Community Call! Come hear what we’re building, ask questions, and connect with the team and other developers in the community. 📅 Thursday, Nov 5 Register here: hubs.la/Q04Dn2GP0

What if you could train a model for your agent harness instead of adapting your harness to the model? OpenHands + @Snowflake Cortex Training let developers post-train open-weight models using their existing agent configurations and tasks. See how it works 👇 openhands.dev/blog/post-trai…

OpenHands is now supported in @scale_AI's open-source AgentEnv framework. Run and evaluate agents against the same tasks and environments, without locking into one agent or infrastructure provider. More interoperability, more choice. That’s the future we’re building toward. Details 👇 agentenvframework.com/docs/a…

Did you miss @gneubig at Hacktoberfest last week? Check out the stream to explore how to move from individual coding agents to repeatable, automated engineering workflows and build an open source software factory 👇 hubs.ly/Q04CY1G80

The cheapest LLM isn’t always the cheapest way to complete a task. New analysis shows how switching models mid-session can erase cost savings through lost context caching. Smart routing needs to optimize the whole workflow, not just model prices. Read the article from @TDataScience: towardsdatascience.com/the-h…

Smarter coding agents aren’t necessarily better at staying within scope. Research highlighted by The New Stack found out-of-scope action rates of 5.4–27.7% for Claude Code, Codex CLI & Gemini CLI vs. 0.2–4.5% for OpenHands. The harness matters.

Open-source AI coding tools are giving developers more choice over models, costs, and how their agents run. Great to see OpenHands featured in this roundup of 5 open-source tools, with a spotlight on autonomous agents running in isolated sandboxes. Check out the full article from @Medium 👇 medium.com/@urano10/%EF%B8%8…

Excited to support for @Recursive_SI Flint 🚀 40B parameters, just 2.5B active per token, built for fast local coding inference. We’ve tested it with OpenHands. Now developers have another open model for real engineering workflows. openhands.dev/blog/recursive…

We’re proud to be a launch partner for @NVIDIA Nemotron 3.5 Super VL, an open multimodal model for agents that see, reason, and act efficiently in specialized workflows, with model control anywhere. The SDLC isn’t text-only. Bugs have screenshots. Incidents have dashboards. Requirements live in docs and diagrams. As agents take on more of the SDLC, they need to understand that context too. The future of agent infrastructure is multimodal + multi-model: the right model for the right work, all within the same agent system. Congrats to the @NVIDIAAI team! 🙌 Read the blog: hubs.la/Q04Bj28z0

OpenHands now integrates with @DigitalOcean Managed Agents! 🌊 Run AI coding agents in isolated cloud environments, pause and resume sessions, and manage everything from Agent Canvas. Use DigitalOcean for agent execution, model inference, or both. More flexibility in where agents run and which models power them. Read more about the integration and how to get started: hubs.la/Q04B3Hbl0

Chat works well when you’re working with one agent. As agent usage grows, we need interfaces built around managing the work, not managing individual conversations. See how OpenHands uses a custom Kanban board to orchestrate agent work: hubs.ly/Q04znB3P0

What’s a task you ask an AI agent to do over and over again? The goal isn’t better prompting. It’s removing the prompt from work that shouldn’t need one. Check out the demo to see how OpenHands turns a repeated prompt into an always-on workflow 👇 hubs.ly/Q04znB1t0

Agentic development isn’t one task for one agent. It’s a system: triage the issue, write the code, run it safely, review the PR, and keep the work moving when feedback comes in. That’s where we focused in September across eight Agent Canvas releases: ⚙️ More powerful automations that carry context from issue → development → PR review 🎛️ Agent Profiles with scoped tools and credentials, so each agent can be configured for the job it needs to do 🐳 Isolated Docker runtimes that give each conversation its own execution environment We also made Apps easier to build and share, alongside dozens of improvements across the Agent Canvas workspace. Read the full September product update: hubs.la/Q04zv2nW0