# SparkForge

SparkForge — an MCP server with tools for screenshots, page scraping, OCR and format conversion. It also generates images and looks up crypto data.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · 219896 installs
- Author: henry-ships
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://smithery.ai/server/henry-ships/sparkforge
- Updated: 2026-03-01
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/mcp/sparkforge

## What it does

SparkForge is a remote MCP server at `https://sparkforge--henry-ships.run.tools`. On the web side, `take_screenshot` captures a page, `scrape_webpage` returns readable text with scripts and tags stripped, `site_crawl` follows a site's internal links, and `get_url_metadata` pulls the title, Open Graph tags and canonical URL. `extract_structured` fills a JSON schema you supply, using Claude AI.

\`convert_format` handles csv to json, json to csv, markdown to html and html to text. `pdf_to_markdown` and `ocr` turn PDFs and images into text. For media there is `generate_image` on Flux (flux-schnell $0.02, flux-dev $0.03, flux-pro $0.05), `generate_video` via fal.ai and `text_to_speech` via ElevenLabs. The crypto tools read CoinGecko, Etherscan (Ethereum and Base wallets) and CryptoPanic.

The PDF and OCR tools take a URL, so a file on your disk can't be passed in. `summarize_text` is algorithmic and makes no AI call; `smart_summarize` is AI-based and has four modes. `list_services` prints the full catalog with prices.

## Requirements
- Remote server: `https://sparkforge--henry-ships.run.tools`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http sparkforge https://sparkforge--henry-ships.run.tools
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add sparkforge --url https://sparkforge--henry-ships.run.tools
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "sparkforge": {
            "url": "https://sparkforge--henry-ships.run.tools"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "sparkforge": {
            "type": "http",
            "url": "https://sparkforge--henry-ships.run.tools"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http sparkforge https://sparkforge--henry-ships.run.tools
```

## When to use it
- get clean text from a webpage without scripts and tags
- convert a PDF at a URL into Markdown
- read text off an image given its URL
- check the title, Open Graph tags and canonical URL of a page
- look up a token's price and the latest crypto headlines

## Example prompts
- Use SparkForge to take a screenshot of this page
- Run `pdf_to_markdown` on the PDF at this link and give me the Markdown
- With SparkForge, pull the title, Open Graph tags and canonical URL from this page

## Similar
- [Arch AI Tools](https://affpapa.org/en/nevibecode/mcp/arch-ai-tools.md) — MCP server · 552 installs
- [Rendex: Rendering API for Images, PDFs & Content Extraction](https://affpapa.org/en/nevibecode/mcp/rendex-rendering-api-for-images-pdfs-content-extraction.md) — MCP server · ★ 6 repo stars
- [mistral-mcp](https://affpapa.org/en/nevibecode/mcp/mistral-mcp.md) — MCP server · ★ 15 repo stars
- [MemeStack.ai](https://affpapa.org/en/nevibecode/mcp/memestackai.md) — MCP server · 2754 installs
- [firecrawl-scraper](https://affpapa.org/en/nevibecode/skills/firecrawl-scraper.md) — skill · 407 installs
- [save-md](https://affpapa.org/en/nevibecode/skills/save-md.md) — skill · 367 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
