Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/groniz/groniz-cli/llms.txt

Use this file to discover all available pages before exploring further.

Claude Code is the primary target for the Groniz agent skill. There are two ways to integrate Groniz with Claude Code: the agent skill, which teaches Claude how to drive the Groniz CLI from the command line, and the MCP server, which exposes Groniz as native tools that Claude can call directly. Both methods work independently, and they can be used together. The agent skill is a SKILL.md file that Claude Code reads from your project’s .claude/skills/ directory. It covers everything groniz --help cannot explain, so Claude can post reliably without trial and error.

​
Install

npx skills add groniz/groniz-cli
Or target Claude Code explicitly to skip the detection prompt: 
npx skills add groniz/groniz-cli -a claude-code
The skill file is installed to .claude/skills/GRONIZ.md inside your project.

​
What Claude learns

Once the skill is installed, Claude understands:
  • Authentication — run groniz auth login once; credentials persist across sessions
  • Media upload — always call groniz upload first and use the returned .path value
  • Channel settings — call groniz integrations settings <id> to discover required fields and limits before composing a post
  • Common traps — strict JSON mode shape, jq path differences, thread chaining with -c, draft validation bypass, and the missing release ID that blocks analytics

​
Try it

After installing the skill and authenticating (see below), try this prompt in Claude Code: 
Schedule a post to my X account saying "Launching today 🚀" at 9am tomorrow
Claude will look up your connected channels, check the integration settings, and construct the correct groniz post command — including any required fields — without you needing to specify flags manually.
Claude Code reads .claude/skills/ automatically when it starts a session in your project directory. No CLAUDE.md entry or other configuration is required.

​
Method 2: MCP server

The Groniz MCP server exposes your Groniz account as tools that Claude can call directly, without going through the CLI. This is useful when you prefer tool-call-based access or are working in an environment where running shell commands is not ideal.

​
Add the MCP server

claude mcp add --transport http https://mcp.groniz.com/mcp -H "Authorization: Bearer YOUR_API_KEY"
Replace YOUR_API_KEY with your Groniz API key. The server uses streamable HTTP transport and requires a valid Bearer token on every request. This exposes Groniz as tools directly inside Claude — Claude can call them without composing shell commands. Requires a paid Claude plan. The MCP integration is not available on free Claude accounts.
The skill and MCP can coexist in the same project. When both are active, the skill helps Claude use the CLI correctly for tasks it handles via the terminal, while MCP handles tool-call-based interactions. You do not need to choose one exclusively.

​
Authenticating the CLI

Before Claude can run any Groniz command, the CLI needs valid credentials. Run this once in your terminal: 
groniz auth login
This stores your API token in ~/.groniz/credentials.json. Claude Code picks up these stored credentials automatically — you do not need to pass an API key on every command. Alternatively, set GRONIZ_API_KEY in your environment for the current session: 
export GRONIZ_API_KEY=your_api_key_here
export only lasts for the current shell session. It does not survive terminal restarts or agent restarts. Use groniz auth login to store credentials persistently in ~/.groniz/credentials.json.

​
Worked example

1

Install the skill

Run the following from your project root:
npx skills add groniz/groniz-cli -a claude-code
The skill is installed to .claude/skills/GRONIZ.md. Commit this file to share the skill with teammates.
2

Authenticate

Run this once in your terminal — you will not need to repeat it:
groniz auth login
Follow the prompts. Your token is saved to ~/.groniz/credentials.json and used automatically by all subsequent groniz commands.
3

Ask Claude to list your channels

Open Claude Code in your project and send:
List my connected social channels
Claude will run groniz integrations list and summarise the results, including each channel’s ID — which you will need for posting.
4

Ask Claude to make a post

Send a natural-language prompt like:
Post "Hello world" to my LinkedIn
Claude will identify your LinkedIn integration ID from the previous step, fetch its settings with groniz integrations settings <id>, compose the post command with the correct flags, and execute it.

Build docs developers (and LLMs) love

Get started for freeTalk to us