Agent37

YC F26 startup selling low-cost agent sandboxes (gVisor or full VM) with prebuilt harnesses for Claude Code, Hermes and opencode.

Agent37 · Infrastructure and inference/serving · infrastructure

Published in Feed

Overview

Agent37 launched on Hacker News on October 12, 2026 with sandboxes for running AI agents, claiming under 10% of typical sandbox-provider cost and about $0.37 per month average spend per sandbox. It offers gVisor sandboxes and full VM sandboxes with Docker, auto-sleep with a cron wake API, prebuilt harness templates (Hermes, OpenClaw, Claude Code, opencode), an open-source OpenAI-compatible gateway and managed MCP connectors. It runs on bare-metal hosts from Hetzner and OVH. It is not SOC 2 certified and gVisor sandboxes lack Docker-in-Docker.

This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.

Key features

  • gVisor sandboxes and full VM sandboxes with Docker
  • Auto-sleep when idle and a cron API to wake for scheduled jobs
  • Harness templates for Hermes, OpenClaw, Claude Code and opencode
  • Open-source OpenAI-compatible gateway; managed MCP connectors and LLM gateway
  • Free plan of roughly 10 awake hours per week

Use cases

  • Running always-on personal or hobby agents cheaply
  • Scheduled agent jobs that sleep between runs

Pricing

Free plan (about 10 awake hours/week); a 2 vCPU / 4 GB box about $4.94/month per founder. Card payment only.

Pricing reflects the source record and may change. Confirm current terms with the provider.

Dates and sources

Released
12 Oct 2026
Record updated
15 Oct 2026
Record last checked
15 Oct 2026

Explore the Feed

Other entries filed under infrastructure and inference/serving; inclusion does not imply equivalent features.

  • NVIDIA PAIR (Personal AI Router)

    NVIDIA's free, open-source personal AI router that spreads local model workloads across PCs on your network; out of beta Oct 13, 2026.

  • Microsoft-Decision-1

    Microsoft's Qwen3.5-9B-based decision-scoring model returns calibrated option probabilities in one pass, at $0.042 per million input tokens.

  • Clef-omni

    Cloudflare's mixture-of-experts decision model that reads text, JSON, images, audio, and video in one request and returns a probability for every allowed answer.

Browse the Feed