Nimply Logo
Gemini CLI logo

Manage social media from Gemini CLI

Google's open-source terminal agent plus Nimply's MCP server: describe the post you want in plain language and Gemini schedules it to your real channels, straight from the command line.

The terminal is where a lot of people already live — scripting, deploying, poking at servers — and Gemini CLI brings an AI agent into that environment rather than asking you to leave it. It reads files, runs commands, and, through MCP, picks up tools for outside services.

Registering Nimply in Gemini CLI's settings takes one JSON block. After that, your social channels are a conversation away from your shell prompt: ask Gemini to check what's scheduled, draft a post from a file in the current directory, queue content to Instagram, Facebook, TikTok, YouTube, LinkedIn, X, Pinterest, or Threads, or report on how last month's posts did.

What is Gemini CLI?

Gemini CLI is Google's open-source command-line AI agent, powered by the Gemini models. It runs as an interactive session in your terminal where the agent can inspect your filesystem, execute shell commands, and reason through multi-step tasks. Support for the Model Context Protocol means servers declared under mcpServers in its settings file become tools the agent can call, checked at any time with the /mcp command.

Nimply's side of the connection is a hosted MCP endpoint — https://mcp.nimply.io/mcp — that Gemini CLI reaches over HTTP with an API key. One detail worth knowing up front: Gemini CLI expects the endpoint under the httpUrl field, not url. The key itself comes from Settings → Developers in your Nimply workspace, is limited to that workspace's permissions, and can be revoked whenever you like.

What Gemini CLI can do with Nimply connected

Post about the thing you just built, from where you built it

You are already in the project directory when the work lands. Gemini can read the release notes or the script output sitting right there and turn it into a scheduled post — no copying text into a browser, no switching mental modes.

Fold publishing into terminal-centric routines

If your day starts and ends in a shell session, your content workflow can too. Ask Gemini for the week's scheduled posts during your morning check-in, or have it queue drafts at the end of the day while you review what shipped.

Pull analytics without opening a dashboard

"Use nimply to list my channels and this month's workspace analytics" gets you a summary in the same window as everything else you monitor. Numbers you would otherwise check occasionally become part of a routine you already run.

Draft in bulk, publish deliberately

Gemini can prepare a batch of drafts across platforms in one session while you keep final publishing manual — or route everything through Nimply's approval workflow so each post gets a human yes before it goes live.

Connect Gemini CLI to Nimply

You need Gemini CLI installed, a Nimply workspace with a connected channel, and an API key from Settings → Developers.

  1. 1

    Create a Nimply API key

    In your Nimply workspace, open Settings → Developers and create an API key. It starts with nim_live_ and only grants access to that one workspace, within the permissions you set.

  2. 2

    Add Nimply to settings.json

    Edit ~/.gemini/settings.json (or the project's .gemini/settings.json for a per-project setup). Use httpUrl for the endpoint — older versions treat url as SSE:

    settings.json
    {
      "mcpServers": {
        "nimply": {
          "httpUrl": "https://mcp.nimply.io/mcp",
          "headers": {
            "Authorization": "Bearer nim_live_YOUR_KEY"
          }
        }
      }
    }
  3. 3

    Restart and verify with /mcp

    Restart the CLI, then run /mcp to confirm the Nimply server is connected and its tools are listed. From there, try: "Use nimply to list my channels and this month's workspace analytics."

    Terminal
    /mcp

Frequently asked questions

The server shows as disconnected in /mcp — what should I check?

First, make sure the endpoint is under httpUrl rather than url; older Gemini CLI versions interpret url as an SSE server and fail to connect. Then confirm the Authorization header is exactly Bearer followed by your nim_live_ key.

Which platforms can I publish to from the terminal?

Every channel connected in your Nimply workspace: Instagram, Facebook, TikTok, YouTube, LinkedIn, X (Twitter), Pinterest, and Threads. Gemini works against the same queue and schedule you would see in the Nimply app.

Is anything installed or hosted on my side for the MCP server?

No. Nimply operates the server at https://mcp.nimply.io/mcp and Gemini CLI connects to it over HTTP. Your settings file holds the URL and your API key, and that is the entire footprint.

Can I use a different Nimply workspace per project?

Yes. Put a workspace-specific key in the project's .gemini/settings.json and keep another in ~/.gemini/settings.json as your default. Each key is bound to exactly one workspace, so the project config determines where posts go.

What stops the agent from publishing something by mistake?

Two layers: your instructions and your key. Tell Gemini to create drafts or request approval rather than publish, and Nimply's approval workflow holds posts for human review. The API key's permissions define the outer limit of what any session can do, and revoking it ends access immediately.

Works well alongside

Ready to try it?

Create a free Nimply account, connect your channels, and you're one config entry away.