Skip to main content

# Agent sandbox

2 related articles

Subscribe to updates

langchain.com Oct 29, 2026
Agents Can Now Click Through the Web Themselves: LangSmith Hosts a Browser and Passes 90% of 100 WebVoyager Tasks
Managed Deep Agents in LangSmith now hosts a browser for agents and supplies four tools for snapshotting, batch actions, screenshots and handoff to a person, so agents can click, scroll and fill forms instead of relying on APIs. On 100 everyday tasks from WebVoyager on live websites, the toolset passed 90% for $1.98 in total, the highest pass rate among seven toolsets tested; the closest, Browserbase's MCP server, passed 89% at 9x the cost. Browser sessions can be scoped per thread or per user, and an agent can hand the browser to a person at a login page and resume afterward.
Managed Deep Agents in LangSmith now hosts a browser for agents and supplies four tools for snapshotting, batch actions, screenshots and handoff to a person, so agents can click, scroll and fill forms instead of relying on APIs. On 100 everyday tasks from WebVoyager on live websites, the toolset passed 90% for $1.98 in total, the highest pass rate among seven toolsets tested; the closest, Browserbase's MCP server, passed 89% at 9x the cost. Browser sessions can be scoped per thread or per user, and an agent can hand the browser to a person at a login page and resume afterward.
Managed Deep Agents Browser use Agent sandbox
Hacker News Oct 13, 2026
Agent37 pushes agent sandboxes to $0.37 a month, funding free tiers by sharing host memory
Agent37 launched low-cost sandboxes for running AI agents, with customers spending an average of $0.37 per sandbox per month and free sandboxes costing roughly $0.05 to $0.10 per user per month in infrastructure. The company began as an OpenClaw wrapper and could not afford a free plan until it scheduled CPU and memory around the bursty usage of agent harnesses, rented hosts from Hetzner and OVH instead of hyperscalers, and used gVisor to share host memory across sandboxes rather than reserving a full VM's memory for each. The tradeoff is that gVisor sandboxes do not support workloads such as Docker-in-Docker or Android emulators, and full VM sandboxes remain available at a slightly higher price.
Agent37 launched low-cost sandboxes for running AI agents, with customers spending an average of $0.37 per sandbox per month and free sandboxes costing roughly $0.05 to $0.10 per user per month in infrastructure. The company began as an OpenClaw wrapper and could not afford a free plan until it scheduled CPU and memory around the bursty usage of agent harnesses, rented hosts from Hetzner and OVH instead of hyperscalers, and used gVisor to share host memory across sandboxes rather than reserving a full VM's memory for each. The tradeoff is that gVisor sandboxes do not support workloads such as Docker-in-Docker or Android emulators, and full VM sandboxes remain available at a slightly higher price.
Agent37 Agent Harness Agent sandbox

You’re all caught up