
1 / 3
【Overview】 exe.dev is a cloud platform built around Linux virtual machines: persistent VMs that you can SSH into as root, and disposable sandboxes that spin up for individual AI agent runs. The persistent VMs come from pools of reserved vCPUs and memory, are served over HTTPS, and support apt and systemd so you can treat them like a normal Linux server. For agent workloads, the platform creates a separate VM per run, each with its own reserved vCPUs and memory, reachable over SSH or HTTPS and billed by the second. The platform also includes Shelley, a web-based coding agent that runs on port 9999 of a default exeuntu VM. Shelley can install software, build a web site, browse the web, and perform other development tasks. Behind the scenes, API tokens carry their own permissions, so you can restrict which commands an agent may run and set when a token expires. Secrets proxying through integrations adds auth headers on the way out, keeping keys on exe.dev rather than inside the VM or the model's context, and connected Google Calendar or Gmail lets an agent add events, send mail as you, and manage labels. This makes exe.dev relevant for solo developers and small apps, teams building at scale, and developers who want a controlled environment for their AI agents. Sharing a running web server or VM works much like sharing a Google Doc, including IAM-based sharing and custom domains. 【Key features】 • Persistent Linux VMs: Reserve a pool of vCPUs and memory, then run persistent Linux VMs with root SSH access, apt, systemd, and HTTPS serving. • Disposable sandboxes for AI agents: Each agent run gets a Linux VM with its own reserved vCPUs and memory, reachable over SSH or HTTPS and billed by the second. • API tokens with command scoping: Tokens carry their permissions, allowing restrictions on which commands an agent may run and when the token expires. • Secrets proxying via integrations: An integration adds the auth header on the way out so keys stay on exe.dev and out of the VM and the model's context. • Shelley coding agent: A web-based coding agent on port 9999 of a default exeuntu VM that can install software, build a web site, browse the web, and more. • Sharing and custom domains: Share a web server or VM with family, friends, and colleagues as easily as a Google Doc, with IAM-based sharing and custom domains. • Google account integration: Connect Google Calendar or Gmail so an agent can add events, send mail as you, and manage labels. 【Use cases】 1. Build apps 2. Run AI agent sandbox sessions 3. Host a web server and share it with family, friends, and colleagues 【Pricing】 Personal: $15/month — Pool size 2 vCPUs / 4 GiB memory, 50 VMs, 100 GiB of filesystem usage included, 200 GiB of network bandwidth. Work: $75/month — Includes 2 vCPUs + 4 GiB to use across your pools, 1,000 VMs per pool, unlimited pools in any region, 100 GiB of filesystem usage included. Enterprise: Custom — Custom resource limits, dedicated hardware and resources, unlimited users; SSO, AWS / GCP VPC support. 【Pros and cons】 Pros: - Developers who want persistent VMs alongside disposable agent sandboxes get both models in one platform, without stitching together separate services. - Small teams benefit from per-second sandbox billing for agent runs, while the flat subscription covers the persistent pool side. - The combination of command-scoped API tokens and secrets proxying gives a practical guardrail for running AI agents without exposing keys. Cons: - The Personal plan's pool is capped at 16 vCPUs / 32 GiB memory per pool, and the Work plan at 512 vCPUs / 1024 GiB per pool, so very large reserves require the Enterprise tier. - Disposable sandboxes are billed at $0.105 per hour per 2 vCPUs / 4 GiB. 【FAQ】 Q: How much does the Personal plan cost, and what does it include? A: The Personal plan is $15/month and includes a pool of 2 vCPUs / 4 GiB memory, 50 VMs, 100 GiB of filesystem usage, and 200 GiB of network bandwidth. Q: Who is exe.dev for? A: It is aimed at solo developers and small apps, teams or those building at scale, and developers running AI agents that need isolated Linux environments. Q: Can I use my own LLM key with Shelley? A: Yes. Shelley uses the LLM integration backed by the LLM Gateway by default, but you can bring your own key.
Personal · $15/month (Pool size 2 vCPUs / 4 GiB memory, 50 VMs, 100 GiB of filesystem usage included) Work · $75/month (Includes 2 vCPUs + 4 GiB to use across your pools, 1,000 VMs per pool, Unlimited pools in any region) Enterprise · Custom (Custom resource limits, Dedicated hardware and resources, Unlimited users; SSO)
WPS AI is an AI assistant built into WPS Office that helps users write documents, generate presentation outlines and…
Baidu Comate is an AI coding assistant built by Baidu on its ERNIE model. It goes beyond simple autocomplete: it…
UniConverter is an AI media conversion and enhancement tool for video, audio, and images. It combines everyday…
6 similar tools to consider instead of exe.dev