# Minterio Mcp Server

> Examples and quick-start for the Minter.io MCP server — connect Claude, ChatGPT, Copilot and Gemini to your Instagram, TikTok, X, Facebook, LinkedIn and Threads analytics.

- **Type:** MCP server
- **Install:** `agentstack add mcp-minter-io-minterio-mcp-server`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [minter-io](https://agentstack.voostack.com/s/minter-io)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [minter-io](https://github.com/minter-io)
- **Source:** https://github.com/minter-io/minterio-mcp-server
- **Website:** https://minter.io/mcp

## Install

```sh
agentstack add mcp-minter-io-minterio-mcp-server
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Minter.io MCP Server — Examples

**Ask your AI assistant about your social media analytics.** Connect **Claude**,
**ChatGPT**, **Copilot** or **Gemini** to your Instagram, TikTok, X/Twitter,
Facebook, LinkedIn and Threads data with the [Minter.io](https://minter.io)
MCP server — connected in under a minute, no API keys to copy.

> **Server endpoint:** `https://minter.io/mcp`
>
> Docs: [minter.io/mcp](https://minter.io/mcp) · Setup guides: [minter.io/connect-to-ai](https://minter.io/connect-to-ai)

Ask things like:

> *"How did my follower growth compare to last month?"*
> *"Which of my posts got the most engagement this quarter?"*
> *"Benchmark my engagement rate against the competitors I track."*

More in the [prompt library](docs/prompts.md).

## What is this?

[Minter.io](https://minter.io) is a social media analytics platform. Its MCP
([Model Context Protocol](https://modelcontextprotocol.io)) server lets any
MCP-capable AI assistant query the reports you already track on Minter.io —
securely, via OAuth 2.0 + PKCE. Your assistant sees exactly the reports you
track, nothing more.

This repo contains:

- Quick-start instructions for Claude, ChatGPT and other assistants
- Example scripts that call the server programmatically (Python & TypeScript)
- A ready-made "weekly report" automation (GitHub Actions)
- A library of [prompts that work well](docs/prompts.md)

## Quick start (no code)

### Claude

1. Open **Settings → Connectors → Add custom connector**
2. Paste `https://minter.io/mcp`
3. Sign in with your Minter.io account when prompted
4. Ask: *"How did my follower growth compare to last month?"*

### ChatGPT

Minter.io is an approved app — add it from the
[ChatGPT app directory](https://chatgpt.com/apps/minter-io/asdk_app_6a2d50ea97c88191b602860a847f8603)
and sign in.

### Other assistants (Copilot, Gemini, …)

Any assistant that supports MCP connectors works the same way: add
`https://minter.io/mcp` as a connector and sign in.

Requires a Minter.io account with the [AI Analyst](https://minter.io/ai-analyst)
feature — included in the [14-day free trial](https://minter.io), no credit
card required.

### Platform-specific setup guides

Step-by-step guides with screenshots for every platform:

| Platform | Claude | ChatGPT |
| --- | --- | --- |
| Instagram | [Guide](https://minter.io/connect-instagram-to-claude) | [Guide](https://minter.io/connect-instagram-to-chatgpt) |
| Instagram hashtags | [Guide](https://minter.io/connect-instagram-hashtags-to-claude) | [Guide](https://minter.io/connect-instagram-hashtags-to-chatgpt) |
| Instagram competitors | [Guide](https://minter.io/connect-instagram-competitors-to-claude) | [Guide](https://minter.io/connect-instagram-competitors-to-chatgpt) |
| TikTok | [Guide](https://minter.io/connect-tiktok-to-claude) | [Guide](https://minter.io/connect-tiktok-to-chatgpt) |
| X/Twitter | [Guide](https://minter.io/connect-twitter-to-claude) | [Guide](https://minter.io/connect-twitter-to-chatgpt) |
| X/Twitter hashtags | [Guide](https://minter.io/connect-twitter-hashtags-to-claude) | [Guide](https://minter.io/connect-twitter-hashtags-to-chatgpt) |
| X/Twitter competitors | [Guide](https://minter.io/connect-twitter-competitors-to-claude) | [Guide](https://minter.io/connect-twitter-competitors-to-chatgpt) |
| Facebook | [Guide](https://minter.io/connect-facebook-to-claude) | [Guide](https://minter.io/connect-facebook-to-chatgpt) |
| Facebook competitors | [Guide](https://minter.io/connect-facebook-competitors-to-claude) | [Guide](https://minter.io/connect-facebook-competitors-to-chatgpt) |
| LinkedIn | [Guide](https://minter.io/connect-linkedin-to-claude) | [Guide](https://minter.io/connect-linkedin-to-chatgpt) |
| Threads | [Guide](https://minter.io/connect-threads-to-claude) | [Guide](https://minter.io/connect-threads-to-chatgpt) |

## Available tools

| Tool | What it does |
| --- | --- |
| `accounts_discovery` | Lists the accounts, hashtags and competitors you track, so the assistant knows what it can analyze |
| `field_discovery` | Shows which metrics are available for a given report (followers, engagement, views, post breakdowns…) |
| `data_query` | Fetches the numbers for any metric and date range |

Call them in that order: `accounts_discovery` → `field_discovery` → `data_query`.

**Notes & limits**

- Dates use `yyyy-mm-dd`; queries default to the last 30 days, up to 2 years of history
- `data_query` returns CSV, max 100 rows per call
- Rate limit: 5000 requests/hour per connection

## Programmatic examples

Both major LLM APIs support remote MCP servers, so you can build your own
automations on top of your Minter.io data with either — the Anthropic
Messages API via the
[MCP connector](https://docs.claude.com/en/docs/agents-and-tools/mcp-connector),
and the OpenAI Responses API via
[MCP tools](https://developers.openai.com/api/docs/guides/tools-connectors-mcp):

| Example | Claude (Anthropic API) | ChatGPT (OpenAI API) |
| --- | --- | --- |
| "Ask a question" CLI, Python | [`ask.py`](examples/python/ask.py) | [`ask_openai.py`](examples/python/ask_openai.py) |
| "Ask a question" CLI, TypeScript | [`ask.ts`](examples/typescript/ask.ts) | [`ask_openai.ts`](examples/typescript/ask_openai.ts) |
| Weekly Markdown report, Python | [`weekly_report.py`](examples/python/weekly_report.py) | — |

- [`.github/workflows/weekly-report.yml`](.github/workflows/weekly-report.yml) —
  runs the weekly report on a schedule and commits it to the repo.
  **Run it in a private repository** — the reports contain your analytics data.

The examples hardcode the model IDs (`claude-sonnet-5` / `gpt-5.6`) and the
Anthropic MCP beta header (`mcp-client-2025-11-20`); when bumping any of these,
update every example.

All examples need:

| Env var | What |
| --- | --- |
| `ANTHROPIC_API_KEY` | Your Anthropic API key (Claude examples) |
| `OPENAI_API_KEY` | Your OpenAI API key (`*_openai` examples) |
| `MINTER_OAUTH_TOKEN` | An OAuth access token for your Minter.io account (obtained via the standard OAuth 2.0 + PKCE flow — see [Authentication](#authentication)) |

## Authentication

The server uses OAuth 2.0 with PKCE (S256) and supports Dynamic Client
Registration, so interactive MCP clients (Claude, ChatGPT, MCP Inspector)
connect automatically — no API key.

For headless/scripted use you complete the OAuth flow once and pass the
resulting access token as a bearer token:

- Authorization endpoint: `https://minter.io/oauth/authorize`
- Protected resource metadata: `https://minter.io/.well-known/oauth-protected-resource`
- Server metadata: `https://minter.io/.well-known/oauth-authorization-server`

The easiest way to grab a token for local experiments is
[MCP Inspector](https://github.com/modelcontextprotocol/inspector):

```bash
npx @modelcontextprotocol/inspector
# connect to https://minter.io/mcp (Streamable HTTP), complete the OAuth flow,
# then copy the access token from the Auth settings
```

> **Note:** access tokens are valid for one year. When a token expires, a
> long-running automation (like the scheduled weekly report) will start
> failing with an authentication error — complete the OAuth flow again and
> update the stored secret.

## Supported platforms

Instagram (accounts, hashtags, competitors) · X/Twitter (accounts, hashtags,
competitors) · TikTok · Facebook (pages, competitors) · LinkedIn (pages) ·
Threads

## Links

- [Minter.io MCP server](https://minter.io/mcp)
- [Connect to AI — setup guides for every platform](https://minter.io/connect-to-ai)
- [AI Analyst](https://minter.io/ai-analyst)
- [Model Context Protocol](https://modelcontextprotocol.io)

## License

Examples in this repo are MIT-licensed. Minter.io itself is a commercial
service — see [minter.io/terms](https://minter.io/terms).

## Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [minter-io](https://github.com/minter-io)
- **Source:** [minter-io/minterio-mcp-server](https://github.com/minter-io/minterio-mcp-server)
- **License:** MIT
- **Homepage:** https://minter.io/mcp

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/mcp-minter-io-minterio-mcp-server
- Seller: https://agentstack.voostack.com/s/minter-io
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.