REST API schema

[alpipego]

Hi, I was wondering if there is a REST API schema somewhere. It seems that at least some parameters are named differently between the python API and the REST API (/crawl), e.g., extraction_strategy vs extraction_config. Somehow, I was unable to find this in the documentation.

[aravindkarnam]

FastAPI(used for REST API server) automatically generates interactive API documentation. If you are running docker container, check under /docs path. If crawl4AI API server is running on port 8000 then check -> "http://localhost:8000/docs".

[aravindkarnam]

@alpipego Yes. We are still testing the new docker setup (as in it's in the repository, but still not fully tested) that why we haven't changed the current image on docker hub, which is more stable. Most like we'll update the image on docker hub within next week.

[alpipego]

There is no JSON schema for what the POST body for a /crawl request should look like, correct?

[aravindkarnam]

@alpipego No, it has body. Looks like this

{
	"urls": [
		"https://docs.crawl4ai.com"
	],
	"browser_config": {
        "type": "BrowserConfig",
        "params": {
            "headless": true,           // Simple boolean - direct value
            "viewport": {               // Complex dictionary - needs type-params
                "type": "dict",
                "value": {
                    "width": 1200,
                    "height": 800
                }
            }
        }
    },
		"crawler_config": {
		"type": "CrawlerRunConfig",
		"params": {
			"deep_crawl_strategy": {
				"type": "BFSDeepCrawlStrategy",
				"params": {
					"max_depth": 2,
					"max_pages":10
				}
			}
		}
     }
}

I'll share a postman collection shortly with different use cases

[alpipego]

Great thanks. I built a little tool to dump the Python objects into JSON and that has gotten me where I need to go meanwhile.

[yoavniran]

hey @aravindkarnam
is that postman collection available somewhere?
It was a real struggle to get the dockerized "crawl" API to work with LLM + Schema extraction ( + openrouter)

The docs and examples in the repo & site are really confusing as they seem to cover past versions and I wasn't able to pinpoint a solid full example for LLM Extraction.

Your example here finally got me through, so thanks!

This is what worked for me in case someone else is struggling:

{
    "urls": [
        "https://website.to.crawl"
    ],
    "browser_config": {
        "type": "BrowserConfig",
        "params": {
            "headless": true, 
            "viewport": {
                "type": "dict",
                "value": {
                    "width": 1200,
                    "height": 800
                }
            }
        }
    },
    "crawler_config": {
        "type": "CrawlerRunConfig",
        "params": {
            "extraction_strategy": {
                "type": "LLMExtractionStrategy",
                "params": {
                    "llm_config": {
                        "type": "LLMConfig",
                        "params": {
                            "provider": "openrouter/deepseek/deepseek-r1:free",
                            "api_token": ""      
                        }
                    },
                    "schema": {
                        "title": "IndexPageLinks",
                        "type": "object",
                        "properties": {
                            "links": {
                                "title": "Links",
                                "type": "array",
                                "description": "List of links found on the page.",
                                "items": {
                                    "type": "string"
                                }
                            }
                        },
                        "required": [
                            "links"
                        ]
                    },
                    "extraction_type": "schema",
                    "instruction": "Scan the content of the page for interesting links. Return a JSON array of links, each url an item in the array..."
                }
            }
        }
    }
}