OpenHelm

The product is managed AI agents that run in isolated cloud sandboxes to automate time-consuming, repeatable work across existing tools and return evidence-backed results for review. It is for developers and business teams in mid-market and enterprise companies who need to delegate tasks such as research, lead enrichment, inbox triage, and data cleanup. It is delivered as a hosted web service with remote MCP server and REST API access and a local desktop app that runs on an existing Claude Code subscription, with all consequential actions held for human sign-off.

Key features

• Isolated cloud sandboxes
• Scoped credential vaults
• Approval queue for consequential actions
• Evidence-backed results for sign-off
• Full audit trail
• Hosted web app
• Remote MCP server connection
• REST API for programmatic control
• Claude Code desktop app
• Scheduled recurring agents
• Live run narration and tracking
• Central workspace for agents and data
• Structured data tables
• Metrics and goals tracking