Failed to call RAG Web Browser: TypeError: fetch failed

[carrot-codeing]

I have the config as follow:

"rag-web-browser": {
                "command": "npx",
                "args": [
                "@apify/mcp-server-rag-web-browser"
                ],
                "env": {
                "APIFY_TOKEN": "apify_api_..."
                }
            }

when i made a call as help me to serach and fecth the content about the GRPO algorithm, it returns the tool calls but error executing, the logs as:

2025-05-17 15:07:19,386 - INFO - 
Assistant: {
    "tool": "search",
    "arguments": {
        "query": "grpo algorithm",
        "maxResults": 5,
        "scrapingTool": "browser-playwright",
        "outputFormats": ["markdown"],
        "requestTimeoutSecs": 60
    }
}
2025-05-17 15:07:25,449 - INFO - Executing tool: search
2025-05-17 15:07:25,450 - INFO - With arguments: {'query': 'grpo algorithm', 'maxResults': 5, 'scrapingTool': 'browser-playwright', 'outputFormats': ['markdown'], 'requestTimeoutSecs': 60} 
2025-05-17 15:07:25,456 - INFO - Executing search...
2025-05-17 15:07:36,095 - WARNING - Error executing tool: Failed to call RAG Web Browser: TypeError: fetch failed. Attempt 1 of 2.
2025-05-17 15:07:36,095 - INFO - Retrying in 1.0 seconds...
2025-05-17 15:07:37,105 - INFO - Executing search...
2025-05-17 15:07:47,681 - WARNING - Error executing tool: Failed to call RAG Web Browser: TypeError: fetch failed. Attempt 2 of 2.
2025-05-17 15:07:47,681 - ERROR - Max retries reached. Failing.
2025-05-17 15:07:47,681 - ERROR - Error executing tool: Failed to call RAG Web Browser: TypeError: fetch failed

Is anything wrong with?

[jirispilka]

Your configuration looks correct. It seems like this may be an intermittent issue.

I tried to replicate it using your queries, and it works on my end.

For the query: "help me to search and fetch the content about the GRPO algorithm"
One of the requests was blocked by Reddit — scraping Reddit can be quite difficult.
If you'd like to exclude Reddit results, you can modify the query like this:
web browser for RAG pipelines -site:reddit.com

For the query: "GRPO algorithm"
There was one result from arXiv (PDF), which we currently aren't able to process properly.

If possible, please raise issues directly at [Apify platform]https://console.apify.com/actors/3ox4R101TgZz67sLr/issues) as well — that way I can check the logs on the platform to see what's going on and whether the Actor was triggered.