← Back to home
Docs

Supported agents

spawnpoint connects to your coding agent over MCP. Every agent below sets up the same way: you ask it to install spawnpoint. Only the one-time sign-in step differs. Technical details (endpoint, transport, auth, tools) are on the MCP server page.

Start here, whichever agent you use. Send your agent this message. It runs the installer, which detects every coding agent on your machine and writes the right config for each. Do not edit an MCP config file by hand: the installer writes the correct settings for each agent, and it is the only setup we test.
Install spawnpoint for yourself: https://getspawnpoint.com/install

To run the installer yourself instead, paste this into a terminal:

curl -fsSL https://getspawnpoint.com/install | bash

On Windows, paste this into PowerShell instead:

irm https://getspawnpoint.com/install.ps1 | iex

Limit it to one agent with SPAWNPOINT_AGENT=claude|codex|cursor|opencode|copilot|vscode|jcode|dsh|antigravity. Then sign in once, in the browser, for the agent you use. The installer prints the step, and your agent passes it on. Find yours below.

Claude Code (terminal)

  1. Ask Claude Code to install spawnpoint (the message above), or run the installer yourself. It also adds the spawnpoint plugin, so "deploy this" works as a skill.
  2. Restart Claude Code.
  3. Run /mcp, choose spawnpoint, choose Authenticate, and approve in the browser.

Claude desktop app

The Claude desktop app includes its own Claude Code runtime that reads the same config as the CLI, so the installer sets it up too, even on a machine that has only the app and no claude command, on a Mac or on Windows. Send the message above in the app's Code tab, then restart the app, open the Code tab again, run /mcp, choose spawnpoint, and Authenticate.

In a chat: add a custom connector

A chat (on claude.ai, or the app's Chat tab) cannot run the installer. Add spawnpoint by hand instead: it takes three clicks and works on every plan (Free allows one custom connector):

  1. Open Settings → Connectors and click Add custom connector.
  2. Name it spawnpoint and paste the URL https://app.getspawnpoint.com/mcp.
  3. Click Add, then Connect and approve the sign-in in your browser.

This route adds spawnpoint to Claude chat and Cowork; the Code tab reads the same connection. You do not get the deploy skill this way (that comes with the plugin the installer adds), but every spawnpoint tool works, and "deploy this" still gets you a link.

Codex CLI

  1. Run the installer. It runs codex mcp add for you.
  2. Run codex mcp login spawnpoint and approve in the browser.

Cursor

  1. Run the installer. It writes the spawnpoint entry into ~/.cursor/mcp.json for you.
  2. Run cursor-agent mcp login spawnpoint, or open Cursor's MCP settings and approve spawnpoint.

Or skip the installer: Add to Cursor adds spawnpoint in one click (Cursor asks to confirm), then approve the sign-in from Cursor's MCP settings.

opencode

  1. Run the installer.
  2. Run opencode mcp auth spawnpoint.

GitHub Copilot CLI

  1. Run the installer.
  2. Run copilot, use any spawnpoint tool, and approve the browser sign-in when it prompts.

VS Code Copilot

  1. Run the installer.
  2. Open Copilot Chat in agent mode, start the spawnpoint server, and sign in from the browser prompt.

Or skip the installer: Install in VS Code adds spawnpoint to VS Code in one click, then start it in Copilot Chat and sign in from the browser prompt.

jcode

  1. Run the installer. It writes the spawnpoint entry into ~/.jcode/mcp.json.
  2. Sign in once:
    npx -y mcp-remote https://app.getspawnpoint.com/mcp
    Approve in the browser, then press Ctrl-C. Every jcode session reuses that sign-in.

jcode supports only stdio MCP today, so the installer connects it to spawnpoint's endpoint through the mcp-remote npm package, which needs Node (npx) on your PATH. The entry sets timeout_secs to 330: jcode cuts an MCP request at 30 seconds by default, and watch_deploy can wait minutes for a slow machine (jcode 0.82 or newer honors it). On a headless machine, create an API token and add "--header" and "Authorization: Bearer spk_live_your_token" to the entry's args instead.

DeepSeek Harness

  1. Run the installer. It adds the spawnpoint entry to ~/.dsh/cordis.patch.yml, the patch layer every dsh profile applies.
  2. Sign in once:
    npx -y mcp-remote https://app.getspawnpoint.com/mcp
    Approve in the browser, then press Ctrl-C. Every dsh session reuses that sign-in.

dsh's MCP client cannot run an OAuth sign-in yet, so the installer connects it to spawnpoint's endpoint through the mcp-remote npm package. The installer changes only the block between its # >>> spawnpoint markers and leaves the rest of the file alone. Sign in before starting dsh: it gives each MCP server a minute at startup, and a headless run that starts without the sign-in has no spawnpoint tools. With no browser, create an API token and add '--header' and 'Authorization: Bearer spk_live_your_token' to the entry's args.

Antigravity (CLI and IDE)

  1. Run the installer. It registers spawnpoint in ~/.gemini/config/mcp_config.json, the file the agy CLI and the IDE share, and adds a short spawnpoint rule to ~/.gemini/GEMINI.md.
  2. CLI: run agy, type /mcp, choose spawnpoint, choose Authenticate, and approve in the browser.
  3. IDE: quit and reopen Antigravity if it was already open (it reads the config when it starts). Open Customizations in the sidebar, choose the Installed tab, and under MCP Servers click Authenticate next to spawnpoint, then approve in the browser. The red "Unauthorized" line shown before that click is the sign-in prompt, not a failure.

Antigravity runs the sign-in itself, with no bridge. The CLI and the IDE each keep their own token, so sign in from the one you use. The rule in GEMINI.md tells the agent that deploying means calling the spawnpoint tools, which keeps it from going to look for a spawnpoint command line. The installer changes only the block between its <!-- >>> spawnpoint --> markers and leaves the rest of the file alone.

Any other MCP client

The installer does not know your client, so this is the one case where you configure it yourself. Point it at https://app.getspawnpoint.com/mcp over Streamable HTTP and complete the OAuth flow it discovers from .well-known. The MCP server page has the exact config format and the per-agent commands for setting up a supported agent by hand.

No browser available (CI, cloud agents)

Clients that cannot do OAuth send a token instead: create a spk_live_… API token in the console and send it as Authorization: Bearer spk_live_…. Both credential types are accepted on the same endpoint, and every call is scoped to the account that authorized it.

Then say "deploy this." After sign-in, every agent works the same way: it makes one tool call and gives you a link. The getting started page covers every step.