We make Pidgo, so we have an interest in this topic. The overview below sticks to what each assistant supports as of November 2026.
MCP, the Model Context Protocol, has become the common way for AI assistants to connect to other apps. Almost every major assistant supports it now. What differs is how you add a server and which kinds of connections each one accepts.
This guide lists the most popular MCP-capable assistants and shows how to connect each to Pidgo, so you can create Instagram comment-to-DM automations by simply describing them. New to MCP? Read What is MCP? first.

Overview
| Assistant | How it connects to Pidgo | Best for |
|---|---|---|
| claude.ai (web and apps) | Sign in | Everyday chat, also on your phone |
| Claude Desktop | Sign in, same connectors as claude.ai | Everyday chat on your computer |
| ChatGPT (web and app) | Sign in, in developer mode | ChatGPT users |
| Claude Code | Sign in or access key | People who work in the terminal |
| Cursor | Sign in or access key | People who already code in Cursor |
| VS Code (GitHub Copilot) | Sign in or access key | VS Code users |
| Gemini CLI | Sign in or access key | Google fans on the terminal |
| OpenAI Codex | Sign in or access key | ChatGPT subscribers who use Codex |
| Windsurf | Access key | Windsurf users |
Every assistant needs the server address https://pidgo.app/api/mcp. Most then open a browser window where you sign in to Pidgo and pick the Instagram account. There's no key to copy, and you can see and disconnect every signed-in assistant in Pidgo under Settings → MCP connection.
If an assistant can't sign in, or you'd rather use a key (for scripts, say), create an access key in the same place and use it as shown under each assistant below.
claude.ai and Claude Desktop
Open Settings → Connectors, choose Add custom connector, name it Pidgo and enter https://pidgo.app/api/mcp. Click Connect, sign in to Pidgo and allow access.
Connectors belong to your Claude account, so Pidgo then shows up in claude.ai, in Claude Desktop and in the Claude apps on your phone.
ChatGPT
Open Settings → Apps & Connectors, turn on Developer mode under Advanced settings, then choose Create. Enter https://pidgo.app/api/mcp as the server address, pick OAuth as the authentication and sign in to Pidgo. In a new chat, turn on Pidgo under the + menu.
Claude Code
Claude Code is Anthropic's assistant for the terminal and the Claude desktop app. Add Pidgo with one command:
claude mcp add --transport http pidgo https://pidgo.app/api/mcp
Then run /mcp in Claude Code, pick pidgo and choose Authenticate. Your browser opens; sign in to Pidgo and allow access. Now ask: "Show me my latest Instagram posts in Pidgo."
With an access key instead, add --header "Authorization: Bearer pdg_..." to the command. Pidgo shows the full command ready to copy when you create a key.
Cursor
Add Pidgo to ~/.cursor/mcp.json, or in Cursor's settings under MCP:
{
"mcpServers": {
"pidgo": {
"url": "https://pidgo.app/api/mcp"
}
}
}
Cursor then asks you to sign in to Pidgo in its MCP settings. With an access key instead, add "headers": { "Authorization": "Bearer pdg_..." } next to the url.
VS Code with GitHub Copilot
Create .vscode/mcp.json in a project, or run MCP: Add Server from the command palette:
{
"servers": {
"pidgo": {
"type": "http",
"url": "https://pidgo.app/api/mcp"
}
}
}
When Copilot first starts Pidgo, VS Code asks you to sign in. Allow it, sign in to Pidgo in the browser, then switch Copilot Chat to agent mode and ask away. With an access key instead, add "headers": { "Authorization": "Bearer pdg_..." } next to the url.
Gemini CLI
Add Pidgo to ~/.gemini/settings.json:
{
"mcpServers": {
"pidgo": {
"httpUrl": "https://pidgo.app/api/mcp"
}
}
}
Then run /mcp auth pidgo in Gemini CLI and sign in to Pidgo in the browser. With an access key instead, add "headers": { "Authorization": "Bearer pdg_..." } next to the httpUrl.
OpenAI Codex
Add Pidgo and sign in from the terminal:
codex mcp add pidgo --url https://pidgo.app/api/mcp
codex mcp login pidgo
With an access key instead, put the key in an environment variable called PIDGO_TOKEN (just the pdg_... part) and add this to ~/.codex/config.toml:
[mcp_servers.pidgo]
url = "https://pidgo.app/api/mcp"
bearer_token_env_var = "PIDGO_TOKEN"
Windsurf
Windsurf connects with an access key. Add Pidgo to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"pidgo": {
"serverUrl": "https://pidgo.app/api/mcp",
"headers": { "Authorization": "Bearer pdg_..." }
}
}
}
Which one should you pick?
- You already use one of them: use that one. They all create the same automations.
- You want the quickest start: claude.ai. Add the address, sign in, done.
- You mostly chat and don't code: claude.ai, Claude Desktop or ChatGPT.
- You manage clients' accounts: any assistant works. Each connection works in one Instagram account, so connect once per client account.
Whichever you choose, the result in Pidgo is the same: a draft automation that you review and switch live yourself. For ideas on what to ask, see 10 AI prompts for Instagram automations.
Questions people ask
Is MCP the same in every assistant?
The protocol is the same, so the same Pidgo tools show up everywhere. What differs is where you add the server and the name of the settings field.
Should I sign in or use an access key?
Sign in wherever your assistant offers it: there's nothing to copy, and you can disconnect each assistant in Pidgo's settings. Use an access key for assistants without sign-in, or for scripts and tools that run on their own.
Do I need a paid plan for my assistant?
That depends on the assistant. Pidgo's MCP connection itself costs nothing extra.
Can I use several assistants at once?
Yes. Each assistant signs in on its own, or gets its own key, so you can disconnect one without affecting the others.
Where do I find my access key again?
Pidgo only shows a key once. If you lose it, create a new one under Settings → MCP connection.