Best MCP Servers for GitHub Copilot in 2026
Firecrawl, Bright Data, Zyte, and Crawlora compared as GitHub Copilot MCP servers — .vscode/mcp.json configs and which fits which job.
GitHub Copilot's agent mode reached MCP support in 2026, and with it the largest existing developer install base of any MCP-capable tool gained access to the same web-data ecosystem that's been growing around Claude Code and Cursor. This guide compares the options that actually connect to Copilot today, with verified .vscode/mcp.json configs — not which vendor markets itself loudest.
How MCP works in GitHub Copilot
VS Code's MCP support (which Copilot's agent mode uses) reads server definitions from .vscode/mcp.json in your workspace, or your global user settings JSON. The root key is servers, not mcpServers — a common copy-paste mistake if you're porting a config from Claude Desktop or Cursor. VS Code also supports an inputs block to prompt for a secret at startup instead of hardcoding it in the file. See Give your AI agent live web data with MCP for how MCP works generally.
What to evaluate
- Generic crawl (any URL → markdown) vs. structured platform data (named sites → JSON).
- Does it need an account/key from request one, or is there a keyless/free path to try it first?
- Tool count and what it actually covers, not just the headline number.
- Does it run as a standard MCP server, or as a Copilot-specific extension that only works inside VS Code?
- Pricing model and whether a failed request still gets billed.
The contenders
| Tool | Format | Coverage | Config | Best for |
|---|---|---|---|---|
| Firecrawl MCP | MCP server | Scrape/crawl/search/extract any URL | .vscode/mcp.json, servers key, npx -y firecrawl-mcp, or the keyless hosted URL mcp.firecrawl.dev/v2/mcp | Generic crawl-to-markdown, try-before-you-key |
| Bright Data | MCP server (60+ tools) | 40+ named platforms + proxy/anti-bot infra | Same .vscode/mcp.json shape, npx @brightdata/mcp or hosted mcp.brightdata.com/mcp?token=... | Anti-bot-hardened extraction at scale |
| Zyte Web Scraping Copilot | VS Code extension (not generic MCP) | Generates a Scrapy project via Copilot Chat | Install from the VS Code Marketplace; requires GitHub Copilot (Pro+ recommended) | A real, deployable Scrapy spider — not a JSON response |
| Crawlora | MCP server | 893 tools across 81 platform groups | .vscode/mcp.json, servers key, https://mcp.crawlora.net/mcp, Authorization: Bearer header | Structured JSON from named platforms, pay-on-success |
Firecrawl MCP — generic crawl, now with a keyless path
Firecrawl's MCP server exposes scrape, crawl, map, search, extract, interact, and its newer agent/monitor/developer_search tools. As of its keyless hosted endpoint (https://mcp.firecrawl.dev/v2/mcp), scrape, search, and interact work with no API key at all (rate-limited) — a genuine try-first-then-key path most competitors don't offer. Full tool access and higher limits still need an Authorization: Bearer <FIRECRAWL_API_KEY> header on the same URL. Choose it when the job is reading or crawling arbitrary websites, and you want to try it before creating an account.
Bright Data — 60+ tools, built for anti-bot scale
Bright Data's VS Code integration is officially documented and explicitly markets itself as giving "GitHub Copilot real-time web search, scraping and structured data tools." Its MCP server exposes 60+ tools across ecommerce, social, business, finance, and browser-automation groups (filterable via a groups= parameter so an agent isn't handed all of them at once), covering named platforms like Amazon, LinkedIn, Instagram, and TikTok. Choose it when reliable extraction from specific, hard, named platforms is the actual constraint.
Zyte Web Scraping Copilot — Copilot writes you a Scrapy spider
Zyte's offering here is genuinely different in shape from the other three: Web Scraping Copilot is a free VS Code extension, not a generic MCP server. It bundles its own local MCP server that Copilot Chat calls to identify page content and generate a real Scrapy spider, with optional deployment to Scrapy Cloud. Requires GitHub Copilot specifically — Copilot Free's rate limits run out quickly, so Pro or better is recommended. This is Copilot's most bespoke integration in this list; it's not portable to other clients the way the other three are. Choose it when you specifically want a maintainable Scrapy codebase generated inside VS Code, and you already pay for Copilot.
Crawlora — structured JSON across 81 platform groups, pay-on-success
Crawlora's MCP server works identically in VS Code/Copilot as in any other MCP client: point .vscode/mcp.json at https://mcp.crawlora.net/mcp with an Authorization: Bearer header, and get 893 tools across search, maps, e-commerce, social, finance, travel, jobs, media, app stores, and reviews — normalized JSON per documented endpoint, not markdown to re-parse. Billing is pay-on-success (a failed request isn't charged), free tier is 2,000 credits/month with no card. Choose it for structured records from named platforms without maintaining a parser.
How to choose, by job
- You need to crawl or read arbitrary websites, and want to try before creating an account → Firecrawl MCP.
- You need reliable extraction from specific hard, named platforms at scale → Bright Data.
- You want Copilot to generate a real, deployable Scrapy project → Zyte Web Scraping Copilot.
- You need structured JSON from named platforms without maintaining a parser, billed only on success → Crawlora MCP.
Add structured web data to GitHub Copilot
893 tools across 81 platform groups, one API key, billed only on success. 2,000 free credits a month, no card.
Sources
Related reading
- Best Web Scraping MCP Servers & Skills for Claude Code in 2026 — the same four vendors, Claude Code's config shape.
- Crawlora vs Firecrawl · Crawlora vs Bright Data · Crawlora vs Zyte.
- Give your AI agent live web data with MCP — how MCP works and how to connect in three minutes.
Frequently asked questions
What's the best MCP server for web scraping with GitHub Copilot?
It depends on the job. Firecrawl MCP fits crawling or reading arbitrary websites, and now offers a keyless hosted endpoint you can try with no API key. Bright Data is built for anti-bot-hardened extraction across 40+ named platforms and is officially documented for VS Code Copilot. Zyte's Web Scraping Copilot is different in shape: it's a Copilot-specific VS Code extension that generates a real Scrapy project rather than exposing generic MCP tools. Crawlora fits when you want structured JSON from named platforms without maintaining a parser, billed only on success.
How do I add an MCP server to GitHub Copilot in VS Code?
Create a .vscode/mcp.json file in your workspace (or add it to your global user settings JSON). The root key is servers, not mcpServers — a common mistake when porting a config from Claude Desktop or Cursor. VS Code also supports an inputs block to prompt for a secret at startup instead of hardcoding an API key in the file.
Can I try Firecrawl MCP without creating an account?
Yes. Firecrawl's hosted endpoint at mcp.firecrawl.dev/v2/mcp works keyless for scrape, search, and interact (rate-limited, no account needed). Full tool access — including crawl, map, and agent — and higher limits need an Authorization: Bearer <FIRECRAWL_API_KEY> header on the same URL.
What does Zyte's Web Scraping Copilot actually do?
It's a free VS Code extension that bundles its own local MCP server, which Copilot Chat calls to identify target content on a page and generate a production Scrapy spider with page objects and a passing test suite — optionally deployable to Scrapy Cloud. It requires GitHub Copilot specifically (Pro or better recommended, since the Free plan's request limits run out quickly) and isn't portable to other MCP clients.
How do I install Crawlora's MCP server for GitHub Copilot?
Add a .vscode/mcp.json with a servers block pointing at https://mcp.crawlora.net/mcp, with an Authorization: Bearer <CRAWLORA_API_KEY> header, using a free API key from crawlora.net (2,000 credits/month, no card).