OpenHelm
Execute one-off or scheduled tasks within a virtual computer that can spin up a real browser, a full coding agent, an email inbox and more.
What it can do
- Run Task: Run any task on OpenHelm by describing it in plain language. OpenHelm spins up a cloud agent with a real browser and your connected tools, performs the work autonomously, and returns a struc
- Run Job: Run a previously created job with concrete inputs. Returns the result if it finishes quickly, otherwise a task_id to retrieve with check_result.
- Create Job: Create a reusable job (task template) whose prompt may contain {{placeholders}}. Run it later with run_job, passing inputs for those placeholders. Useful when you'll repeat the same task w
What data it sees
Do you need an account
No: the server works without sign-in
Execute one-off or scheduled tasks within a virtual computer that can spin up a real browser, a full coding agent, an email inbox and more.
OpenHelm unlocks the ability for your agent to carry out all sorts of tasks, such as: • Deep research on a topic • Emailing + handling responses • Coding work • Updating files
What makes OpenHelm the best solution for this: • Underpinned by a powerful AI that monitors work and ensures goals are actually achieved • Browser is SUPER stealthy, meaning it can access more of the web than other bots • Work is managed by a powerful agentic system, facilitating multi-step flows
For performing complex, one-off or recurring work, OpenHelm provides a secure, remote environment to make it all possible.
Server tool list (10)
Raw names from tools/list. Only developers need these.
| run_task | Run any task on OpenHelm by describing it in plain language. OpenHelm spins up a cloud agent with a real browser and your connected tools, performs the work autonomously, and returns a structured result with evidence (screenshot + matched text), a confidence level, and usage/cost. Use this for one-off research, web automation, data extraction, monitoring, or any goal you'd give a capable operator. Pass an `output_schema` to get back exactly the fields you need. Returns the result if the task finishes quickly, otherwise a task_id you retrieve with check_result. |
| run_job | Run a previously created job with concrete inputs. Returns the result if it finishes quickly, otherwise a task_id to retrieve with check_result. |
| create_job | Create a reusable job (task template) whose prompt may contain {{placeholders}}. Run it later with run_job, passing inputs for those placeholders. Useful when you'll repeat the same task with different parameters. |
| create_project | Create a new OpenHelm project (a container for jobs and runs). Returns the project id. |
| list_projects | List the OpenHelm projects this key can access. Projects are the containers tasks run in. |
| list_runs | List recent API runs for this key, most recent first. Optionally filter by job_id. |
| list_connections | List the connections (credentials and tools) available to this account — Gmail, GitHub, MCP servers, etc. — with their auth status. Tasks that need a connection use whatever is set up here; if a run fails with failure_reason 'auth_required', the needed connection isn't set up yet. Secrets are never returned. |
| create_connection | Create a connection so your tasks can authenticate. Two kinds: a 'token' connection stores an API key/secret (e.g. a GitHub PAT) in encrypted storage; an 'mcp' connection registers a remote HTTP/SSE MCP server by URL with optional headers. For OAuth providers that need a browser (Google, Notion, …) use create_connection_link instead. The secret is write-only. |
| create_connection_link | Get a short-lived link to connect an OAuth or browser-login provider (Google, Notion, LinkedIn, …) that can't be set up from a raw key. Returns a URL the user opens once to authorise; then poll list_connections until it shows as authenticated. |
| check_result | Retrieve the result of a previously started OpenHelm task. Use this when an earlier tool returned status "running" together with a task_id — call check_result with that task_id to get the finished result (or the current status if it's still running). Returns the structured result, confidence, evidence (screenshot URL + matched text), and usage when the task has succeeded. |