Skip to content
MCP server

agend-sh

Persistent Linux environments for AI agents, with interactive terminals and HTTPS previews

Runs locallyStatus and history
Pricing
Free trial
Platforms
Linux, CLI, API
Runs
Locally
Version
1.2.14
Listing updated
1 day ago

About agend-sh

agend-sh is an MCP server for cloud and infrastructure. It runs locally on your machine and is started by your AI client.

agend gives an agent its own isolated Linux machine to drive over MCP. The headline feature is input readiness: when a program in the terminal stops and waits for typing, such as an installer prompt, a Python REPL or a database session, the tool response says so, and the agent can answer in the same session. Sessions and files persist between tool calls, and an agent can start services and open a browser preview URL, or you can mirror its terminal with agend watch.

Each environment runs in a Firecracker microVM with its own guest kernel. Idle environments sleep, after 30 minutes on Developer and 2 hours on Pro, and external connections may need to reconnect after waking. The site is clear that task success still needs its own verification, and that if an API or plain script already does the job, you should keep using it.

You bring your own agent and model; agend charges for compute and storage. A 7-day trial needs no card, paid plans start at $39 per month, and it requires an account and environment quota. The CLI installs by script or Homebrew and can configure Claude Code, Codex, Cursor and Gemini.

Install

Run in a terminal

claude mcp add agend-sh --env AGEND_API_TOKEN='<AGEND_API_TOKEN>' -- docker run -i --rm -e AGEND_API_TOKEN ghcr.io/agend-sh/agend-mcp@sha256:4462e2571be0e69003e1be93255e04523342e1e5aa17482ec3d63bd7506bb82f

Add to claude_desktop_config.json

{
  "mcpServers": {
    "agend-sh": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "AGEND_API_TOKEN",
        "ghcr.io/agend-sh/agend-mcp@sha256:4462e2571be0e69003e1be93255e04523342e1e5aa17482ec3d63bd7506bb82f"
      ],
      "env": {
        "AGEND_API_TOKEN": "<AGEND_API_TOKEN>"
      }
    }
  }
}

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "agend-sh": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "AGEND_API_TOKEN",
        "ghcr.io/agend-sh/agend-mcp@sha256:4462e2571be0e69003e1be93255e04523342e1e5aa17482ec3d63bd7506bb82f"
      ],
      "env": {
        "AGEND_API_TOKEN": "<AGEND_API_TOKEN>"
      }
    }
  }
}

Add to .vscode/mcp.json

{
  "servers": {
    "agend-sh": {
      "type": "stdio",
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "AGEND_API_TOKEN",
        "ghcr.io/agend-sh/agend-mcp@sha256:4462e2571be0e69003e1be93255e04523342e1e5aa17482ec3d63bd7506bb82f"
      ],
      "env": {
        "AGEND_API_TOKEN": "<AGEND_API_TOKEN>"
      }
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "agend-sh": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "AGEND_API_TOKEN",
        "ghcr.io/agend-sh/agend-mcp@sha256:4462e2571be0e69003e1be93255e04523342e1e5aa17482ec3d63bd7506bb82f"
      ],
      "env": {
        "AGEND_API_TOKEN": "<AGEND_API_TOKEN>"
      }
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "agend-sh": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "AGEND_API_TOKEN",
        "ghcr.io/agend-sh/agend-mcp@sha256:4462e2571be0e69003e1be93255e04523342e1e5aa17482ec3d63bd7506bb82f"
      ],
      "env": {
        "AGEND_API_TOKEN": "<AGEND_API_TOKEN>"
      }
    }
  }
}

Configuration

  • AGEND_API_TOKENYour Agend account token. Sign in with agend login, then follow docs/mcp-registry.md to load the active account token. An Agend account and available environment quota are required.Required

Tools this server provides (3)

Taken from the project documentation; the running server may differ.

list_environments
List your agend environments
shell_exec
Run a command, optionally interactive
shell_send_raw
Send raw input to an interactive session

What it does

Input-readiness signals
Responses report when a terminal program is waiting for input.
Persistent sessions
REPL and database state survives between tool calls.
Firecracker isolation
Each environment is a microVM with its own kernel.
Interactive tools
Send keystrokes and read output through shell_exec and shell_send_raw.
Preview URLs
Expose services over HTTPS and watch the terminal live.
Client setup
The CLI configures several MCP clients automatically.

Good for

  • Let an agent answer an installer's interactive prompts
  • Keep a Python variable alive across tool calls
  • Run a CLI test suite that needs repeated navigation
  • Preview a web app an agent is building

Pricing

Free trial, from $39 a month

7-day trial, no card; plans from $39/month

Prices change. Check agend.sh before you buy.

agend-sh FAQ

Is there a trial?

Yes, 7 days with one small environment and no credit card.

What do plans cost?

Plans start at $39 per month; they charge for compute and storage, not AI usage.

Does it include a model?

No. You bring your own agent and model access.

What isolates the environments?

Each runs in a Firecracker microVM with its own guest kernel.

Is agend-sh free?

agend-sh is a paid product with a free trial. Paid plans start at $39 a month.

How do I add agend-sh to Claude Code?

Run this command in your terminal: claude mcp add agend-sh --env AGEND_API_TOKEN='<AGEND_API_TOKEN>' -- docker run -i --rm -e AGEND_API_TOKEN ghcr.io/agend-sh/agend-mcp@sha256:4462e2571be0e69003e1be93255e04523342e1e5aa17482ec3d63bd7506bb82f

What are the best alternatives to agend-sh?

The closest alternatives to agend-sh are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

Alternatives to agend-sh

  1. REA

    MCP server

    CLI and MCP server that lets coding agents reverse engineer apps and binaries

    Runs locally
    • Open source
    • 464K GitHub stars
    • Developer tools
  2. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  3. GitHub

    MCP server

    Official MCP server connecting AI assistants to GitHub repos, issues and workflows

    Not checked yet
    • Remote
    • Open source
    • 34.5K GitHub stars
    • Developer tools
  4. MCP toolkit giving coding agents symbol-level code retrieval and editing

    Runs locally
    • Open source
    • 31.5K GitHub stars
    • Developer tools
  5. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  6. Ouroboros

    MCP server

    Spec-first layer that pins requirements and verifies results of AI coding agent work

    Runs locally
    • Open source
    • 6.4K GitHub stars
    • Developer tools