firecrawl_scrape : Status code: 400 : When 'extract' or 'json' format is specified....

[bdb123]

Hi, getting this error :

Failed to scrape URL. Status code: 400. Error: Bad Request - [{"code":"custom","message":"When 'extract' or 'json' format is specified, corresponding options must be provided, and vice versa","path":[]}]

I'm using MCP Inspector to run and connect to the local firecrawl mcp server. This error only seems to happen for scrape. Map and crawl are working OK.

I'm asking for markdown only so not asking for json or extract. I've tried playing around with other parameters and checking against the settings in a successful playground execution but not getting anywhere. So either I'm missing something or its a wee bug... help would be much appreciated as this is such a useful tool and I really want to evaluate and get it into an agent!

Request to server:
{
"method": "tools/call",
"params": {
"name": "firecrawl_scrape",
"arguments": {
"url": "bbc.co.uk",
"formats": [
"markdown"
],
"includeTags": [],
"excludeTags": [],
"actions": [],
"extract": {},
"location": {}
},
"_meta": {
"progressToken": 6
}
}
}

Response:
{
"content": [
{
"type": "text",
"text": "Failed to scrape URL. Status code: 400. Error: Bad Request - [{"code":"custom","message":"When 'extract' or 'json' format is specified, corresponding options must be provided, and vice versa","path":[]}]"
}
],
"isError": true
}

Thanks!!

[JordanCason]

Im also getting a 404

firecrawl service is up and working, also works with my other tools:

curl -X POST http://localhost:3002/v1/scrape \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer NOT_NEEDED_CURRENTLY" \
  -d '{
    "url": "https://docs.firecrawl.dev",
    "formats": ["markdown", "html"]
  }'

However MCP server never even seems to hit the firecrawl service and returns a 404
npx @modelcontextprotocol/inspector npx -y firecrawl-mcp -e FIRECRAWL_API_URL=http://localhost:3002/ -e FIRECRAWL_API_KEY=NOT_NEEDED_CURRENTLY

Initializing Firecrawl MCP Server...
Running in stdio mode, logging will be directed to stderr
[info] Firecrawl MCP Server initialized successfully
[info] Configuration: API URL: http://localhost:3002/ Firecrawl MCP Server running on stdio
[info] [2025-05-13T10:46:17.960Z] Received request for tool: firecrawl_scrape 
"Unexpected error occurred while trying to scrape URL. Status code: 404"

[bdb123]

I'm on windows and this opens the inspector OK : npx @modelcontextprotocol/inspector npx -y firecrawl-mcp

then I set transport to STDIO and manually add an environment variable FIRECRAWL_API_KEY with my api key. This connects OK. And I'm able to list available tools. But I get my 400 when i try run firecrawl_scrape. I was able to succesfully run other commands and saw the results in the firecrawl history so connections sucessfull for sure and my 400 is a server response i think, jsut tried it now and those others arent working :-/ but I need to move on and find workarounds :-)

I really am no expert but I think you don't need the URL..... I think the defaults all work as inspector is effectively creating a sub process, executing the npx for firecrawl in that, then using stdio for for interprocess transport.... could be speaking rubbish, still trying to understand MCP fully and the tooling around it. If you are using SSE as the transport then you would need a URL to target.

[zhangluustb]

Hi, getting this error :

Failed to scrape URL. Status code: 400. Error: Bad Request - [{"code":"custom","message":"When 'extract' or 'json' format is specified, corresponding options must be provided, and vice versa","path":[]}]

I'm using MCP Inspector to run and connect to the local firecrawl mcp server. This error only seems to happen for scrape. Map and crawl are working OK.

I'm asking for markdown only so not asking for json or extract. I've tried playing around with other parameters and checking against the settings in a successful playground execution but not getting anywhere. So either I'm missing something or its a wee bug... help would be much appreciated as this is such a useful tool and I really want to evaluate and get it into an agent!

Request to server: { "method": "tools/call", "params": { "name": "firecrawl_scrape", "arguments": { "url": "bbc.co.uk", "formats": [ "markdown" ], "includeTags": [], "excludeTags": [], "actions": [], "extract": {}, "location": {} }, "_meta": { "progressToken": 6 } } }

Response: { "content": [ { "type": "text", "text": "Failed to scrape URL. Status code: 400. Error: Bad Request - [{"code":"custom","message":"When 'extract' or 'json' format is specified, corresponding options must be provided, and vice versa","path":[]}]" } ], "isError": true }

Thanks!!

same problem. it's a bug. have you solved it?

[bdb123]

Nope, next step was to step through the MCP code and see if I can work it out, that looks doable. But frankly, as a python newbie with a ton of deliverables, it would be a rabbit hole for me right now. I'm still using firecrawl but just the API direct so MCP/Agent experimentation will have to be later when this is a bit more mature.