Run a Firecrawl agent and toolkit in the cloud using Meshagent:

​
1. Set Required Environment Variables

Configure the following environment variables for Firecrawl:

  • FIRECRAWL_API_URL: URL for Firecrawl API (e.g., https://api.firecrawl.dev/v1)
  • FIRECRAWL_RETRY_MAX_ATTEMPTS: Max retry attempts for API calls (e.g., 5)
  • FIRECRAWL_RETRY_INITIAL_DELAY: Initial delay before retry in ms (e.g., 2000)
  • FIRECRAWL_RETRY_MAX_DELAY: Maximum delay before retry in ms (e.g., 30000)
  • FIRECRAWL_RETRY_BACKOFF_FACTOR: Exponential backoff factor for retries (e.g., 3)
  • FIRECRAWL_CREDIT_WARNING_THRESHOLD: Credits remaining to begin warning (e.g., 2000)
  • FIRECRAWL_CREDIT_CRITICAL_THRESHOLD: Credits at critical usage (e.g., 500)
  • FIRECRAWL_API_KEY: Your Firecrawl API key

​
2. Install Meshagent

pip install "meshagent[all]"

​
3. Sign up and Authenticate

Follow the guide at Meshagent CLI: Get Started to create an account and authenticate.

​
4. Run the Firecrawl MCP Server in a Meshagent Room

meshagent service test \
  --room=test \
  --role=agent \
  --image=meshagent/mcp-firecrawl:latest \
  --env MESHAGENT_PORT=8001 \
  --env FIRECRAWL_API_URL=https://api.firecrawl.dev/v1 \
  --env FIRECRAWL_RETRY_MAX_ATTEMPTS=5 \
  --env FIRECRAWL_RETRY_INITIAL_DELAY=2000 \
  --env FIRECRAWL_RETRY_MAX_DELAY=30000 \
  --env FIRECRAWL_RETRY_BACKOFF_FACTOR=3 \
  --env FIRECRAWL_CREDIT_WARNING_THRESHOLD=2000 \
  --env FIRECRAWL_CREDIT_CRITICAL_THRESHOLD=500 \
  --env FIRECRAWL_API_KEY=YOUR-API-KEY \
  --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" \
  --name=mcp-firecrawl-service
  • This starts a Firecrawl MCP server as a service accessible in the test Meshagent Room. (Rooms will close if they go inactive.)

​
5. Join the Room with a Chatbot Agent (Using the Firecrawl Toolkit)

meshagent chatbot join \
  --room=test \
  --agent-name=sample \
  --name=sample \
  --toolkit=mcp-firecrawl
  • This launches a chatbot in the Meshagent room that can utilize the Firecrawl tools.
  • You can use multiple toolkits per agent, and the command output will include a link to interact in the Meshagent room.

​
Tools Available

Below are the Firecrawl tools provided when using this MCP server in Meshagent Rooms.

​
Tool Summary

ToolShort Description
firecrawl_check_crawl_statusCheck the status of a crawl job.
firecrawl_crawlStarts an asynchronous crawl job on a website and extracts content from all pages.
firecrawl_deep_researchConduct deep web research on a query using intelligent crawling, search, and LLM analysis.
firecrawl_extractExtract structured information from web pages using LLM capabilities.
firecrawl_generate_llmstxtGenerate a standardized llms.txt (and optionally llms-full.txt) file for a given domain.
firecrawl_mapMap a website to discover all indexed URLs on the site.
firecrawl_scrapeScrape content from a single URL with advanced options.
firecrawl_searchSearch the web and optionally extract content from search results.

​
Tools Details

​
firecrawl_check_crawl_status

Check the status of a crawl job.

Parameters:

  • id (string): Crawl job ID to check.

Returns: Status and progress of the crawl job, including results if available.

​
firecrawl_crawl

Starts an asynchronous crawl job on a website and extracts content from all pages.

Parameters:

  • url (string): Starting URL for the crawl
  • (Optional) Other controls: allowBackwardLinks, allowExternalLinks, deduplicateSimilarURLs, excludePaths, ignoreQueryParameters, ignoreSitemap, includePaths, limit, maxDepth, scrapeOptions, webhook

Returns: Operation ID for status checking; use firecrawl_check_crawl_status to check progress.

​
firecrawl_deep_research

Conduct deep web research on a query using intelligent crawling, search, and LLM analysis.

Parameters:

  • query (string): The query to research
  • (Optional) maxDepth, maxUrls, timeLimit

Returns: LLM-generated analysis from deep research, including sources.

​
firecrawl_extract

Extract structured information from web pages using LLM capabilities.

Parameters:

  • urls (array): List of URLs to extract info from
  • (Optional) allowExternalLinks, enableWebSearch, includeSubdomains, prompt, schema, systemPrompt

Returns: Extracted structured data as defined by your schema.

​
firecrawl_generate_llmstxt

Generate a standardized llms.txt (and optionally llms-full.txt) file for a given domain.

Parameters:

  • url (string): The URL to generate LLMs.txt from
  • (Optional) maxUrls, showFullText

Returns: LLMs.txt and optionally llms-full.txt file contents.

​
firecrawl_map

Map a website to discover all indexed URLs on the site.

Parameters:

  • url (string): Starting URL for URL discovery
  • (Optional) ignoreSitemap, includeSubdomains, limit, search, sitemapOnly

Returns: Array of URLs found on the site.

​
firecrawl_scrape

Scrape content from a single URL with advanced options.

Parameters:

  • url (string): The URL to scrape
  • (Optional) actions, excludeTags, extract, formats, includeTags, location, mobile, onlyMainContent, removeBase64Images, skipTlsVerification, timeout, waitFor

Returns: Content in the requested format (e.g., Markdown, HTML).

Search the web and optionally extract content from search results.

Parameters:

  • query (string): Search query string
  • (Optional) country, filter, lang, limit, location, scrapeOptions, tbs

Returns: Array of search results (with optional scraped content).

​
Additional Resources