firecrawl_search MCP tool is missing documented /v2/search fields

[ethanaggor]

Describe the Bug

The firecrawl_search MCP tool schema does not expose all documented public /v2/search request fields.

Checked against firecrawl/firecrawl-mcp-server origin/main at 400072e.

Current firecrawl_search schema exposes:

{
  query: string;
  limit?: number;
  tbs?: string;
  filter?: string;
  location?: string;
  sources?: Array<{ type: "web" | "images" | "news" }>;
  scrapeOptions?: ...;
  enterprise?: Array<"default" | "anon" | "zdr">;
}

The public /v2/search API also documents:

{
  categories?: Array<"github" | "research" | "pdf" | { type: "github" | "research" | "pdf" }>;
  country?: string;
  timeout?: number;
  ignoreInvalidURLs?: boolean;
}

These fields are not currently available through the MCP tool schema.

To Reproduce

1. Configure the Firecrawl MCP server.
2. Inspect the firecrawl_search tool schema.
3. Try to call firecrawl_search with documented /v2/search fields such as:

{
  "query": "machine learning",
  "categories": ["research"],
  "limit": 5
}

or:

{
  "query": "restaurants",
  "country": "DE",
  "location": "Germany",
  "limit": 5
}

or:

{
  "query": "Firecrawl docs",
  "timeout": 30000,
  "ignoreInvalidURLs": true,
  "limit": 5
}

4. These fields are not accepted by the current MCP tool schema.

Expected Behavior

The MCP firecrawl_search tool should expose the documented /v2/search request fields.

Expected schema shape:

{
  query: string;
  limit?: number;
  sources?: Array<
    | "web"
    | "images"
    | "news"
    | { type: "web" | "images" | "news" }
  >;
  categories?: Array<
    | "github"
    | "research"
    | "pdf"
    | { type: "github" | "research" | "pdf" }
  >;
  tbs?: string;
  location?: string;
  country?: string;
  timeout?: number;
  ignoreInvalidURLs?: boolean;
  filter?: string;
  enterprise?: Array<"default" | "anon" | "zdr">;
  scrapeOptions?: ...;
}

The implementation should forward accepted MCP fields to /v2/search.

Screenshots

Not applicable.

Environment

• Repository: firecrawl/firecrawl-mcp-server
• Ref checked: origin/main at 400072e
• Package version on main: 3.13.0
• Published npm latest checked: firecrawl-mcp@3.11.0
• Node.js Version: Not version-specific; schema/source inspection. Local verification used Node v25.2.1.

Logs

Not applicable.

Additional Context

The current README Search example includes country:

{
  "name": "firecrawl_search",
  "arguments": {
    "query": "latest AI research papers 2023",
    "limit": 5,
    "country": "us",
    "sources": [
      { "type": "web" },
      { "type": "images" },
      { "type": "news" }
    ]
  }
}

But country is not present in the actual MCP tool schema.

The public /v2/search docs also document categories, country, timeout, and ignoreInvalidURLs.

The MCP implementation currently calls the Node SDK client.search(). The SDK currently does not forward enterprise, country, or expose the full Search response metadata.

Related SDK issues:

• [Bug] Node SDK v2 search does not forward all documented /v2/search request fields firecrawl#3437
• [Feat] Expose full /v2/search response metadata in Node SDK firecrawl#3438