Why does firecrawl-mcp-server exit silently when I set SSE_LOCAL=true?
I ran firecrawl-mcp-server with SSE_LOCAL=true and it exited immediately without printing port or SSE endpoint information. How should SSE_LOCAL=true be used now, and is there a working way to run it for MCP clients?
Originally asked on GitHub: https://github.com/firecrawl/firecrawl-mcp-server/issues/47
4 answers
Status: still open, with no maintainer response in the thread. A later comment reports that SSE_LOCAL appears to have been deprecated or changed in recent versions of the npm package. Instead of the old messages ("Starting server on port 3000", "MCP SSE Server listening on http://localhost:3000", "SSE endpoint: http://localhost:3000/sse", "Message endpoint: http://localhost:3000/messages"), the current build with SSE_LOCAL=true only prints something like: [INFO] 2025-09-18T22:58:32.126Z [FastMCP info] Starting server in stateless mode on HTTP Stream at http://localhost:3000/mcp. So the process may actually be starting, but the expected SSE endpoint output is gone. One commenter confirmed that running the server in Docker worked for them with this exact command: docker run -d -p 8000:3000 --name firecrawl -e FIRECRAWL_API_URL=http://XXX -e SSE_LOCAL=true dockerpull.cn/mcp/firecrawl. That command maps host port 8000 to container port 3000. If your local process still exits immediately, run it in the foreground and check the logs to see whether FIRECRAWL_API_URL is reachable; that last part is general advice, not something confirmed in this thread.
Did this work for you?What a GitHub user tried on GitHub (Didn't work)
The reporter ran the command with SSE_LOCAL=true locally and observed no monitoring or port output before the process ended.
Source: https://github.com/firecrawl/firecrawl-mcp-server/issues/47 (opens external site)
Did this work for you?What @Vincentdu-cn tried on GitHub (Worked)
A commenter ran firecrawl as a Docker container using docker run -d -p 8000:3000 --name firecrawl -e FIRECRAWL_API_URL=http://XXX -e SSE_LOCAL=true dockerpull.cn/mcp/firecrawl and reported it worked fine.
Did this work for you?What @ddalton15 tried on GitHub (Outcome unclear)
A later comment observed that the npm build with SSE_LOCAL=true logs only the FastMCP stateless HTTP Stream startup line at http://localhost:3000/mcp instead of the old SSE listening messages, suggesting SSE_LOCAL may be deprecated.
Did this work for you?
Related questions
- How do I make firecrawl-mcp-server use a self-hosted Firecrawl URL?
- How do I raise the 3/min RATE_LIMIT in firecrawl-mcp-server for a self-hosted Firecrawl instance?
- firecrawl_check_batch_status returns a localhost URL that doesn't work in Docker — how to get batch results?
- How to fix firecrawl-mcp hanging on stdio during init and tool calls?
- Firecrawl MCP server not deployed on Smithery after repo moved to Mendable