urlsarrayrequired
URLs to extract content from.
https://api.sandbase.ai/v1/api/tavily/extractExtracts clean text content from specified web pages using their URLs. Supports basic or advanced extraction depth and optional image inclusion.
{
"failed_results": [],
"request_id": "35001ccb-b475-4057-b008-b5026dc5ac41",
"response_time": 0,
"results": [
{
"images": [],
"raw_content": "# Example Domain\n\nThis domain is for use in documentation examples without needing permission. Avoid use in operations.\n\n[Learn more](https://iana.org/domains/example)",
"title": "Example Domain",
"url": "https://example.com"
}
]
}Select an example to load its inputs and result into the playground.
Returns the page title and images for each URL you send, with a list of URLs that could not be extracted. It needs a urls array with one or more web page addresses.
include_images to true.failed_results.curl -X POST "https://api.sandbase.ai/v1/api/tavily/extract" \
-H "Authorization: Bearer $SANDBASE_API_KEY" \
-H "Content-Type: application/json" \
-d '{"urls": ["https://en.wikipedia.org/wiki/Artificial_intelligence"]}'
The result is returned in outputs[0].data.
| Name | Type | Required | Description |
|---|---|---|---|
urls | array | Yes | URLs to extract content from. |
extract_depth | string | No | Depth of content extraction. |
include_images | boolean | No | Include images found on the pages. |
Real response for the request above, trimmed: lists keep their first item and long strings are cut.
{
"failed_results": [],
"request_id": "f594f392-10f6-4a3f-9d7c-ac020a617625",
"response_time": 0.03,
"results": [
{
"images": [],
"title": "Artificial intelligence - Wikipedia",
"url": "https://en.wikipedia.org/wiki/Artificial_intelligence"
}
],
"usage": {
"credits": 1
}
}
| Field | Meaning |
|---|---|
results[] | One entry per URL that was extracted successfully. |
results[].url | The page address that was extracted. Use it to match each result to the URL you sent. |
results[].title | The title of the page, for example Artificial intelligence - Wikipedia. |
results[].images | List of images found on the page. It is empty in this response, where include_images was not set. |
failed_results | List of URLs that could not be extracted. It is empty when every URL worked. |
request_id | Unique ID of the request, as a UUID. |
response_time | Time taken for the request. The value is a number such as 0.03. |
usage.credits | Number of credits recorded for the call, such as 1. |
urls array to extract them in one call. Check both results and failed_results for each one.extract_depth to basic or advanced. Set include_images to true to fill results[].images.Send urls as an array of public page addresses. extract_depth accepts basic or advanced; include_images is a boolean. Search takes a query, while this request takes URLs you already selected. The current reference sample shows URL, title and image fields with failed_results; it does not demonstrate raw_content. That absence is not proof that text extraction is impossible, and it is not a guarantee of full-page text either. Check the approved response for your target before writing a text parser or claiming a complete RAG workflow.
To discover candidate URLs first, read the Tavily Search parameters and the Python search guide. Review and select a URL yourself; the examples do not chain or retry requests.
$0.012 per successful call. Failed calls are not charged. It is included in the daily free quota of each organization.
Start building
Platform API endpoint with unified authentication and usage tracking.
# The quoted heredoc keeps Unicode and shell metacharacters unchanged.
result=$(curl --fail-with-body --silent \
-X POST "https://api.sandbase.ai/v1/api/tavily/extract" \
-H "Authorization: Bearer $SANDBASE_API_KEY" \
-H "Content-Type: application/json" \
--data-binary @- <<'SANDBASE_JSON'
{
"urls": [
"https://static.sandbase.ai/examples/de3a14cf-48e2-4dd0-9ef9-4c4c1da485c8/input/2f4290fd-6946-4f6b-8cbe-5b3366e41080.com"
]
}
SANDBASE_JSON
)
printf '%s\n' "$result"Set your SandBase API key in your server environment. Requests use Bearer authentication.
export SANDBASE_API_KEY="YOUR_API_KEY"Keep your key on the server. Do not expose it in browser code.
urlsarrayrequiredURLs to extract content from.
extract_depthstringDepth of content extraction.
Possible values: basic, advanced
include_imagesbooleanInclude images found on the pages.
Saved example response for this API; no live request is made when viewing it.
{
"failed_results": [],
"request_id": "35001ccb-b475-4057-b008-b5026dc5ac41",
"response_time": 0,
"results": [
{
"images": [],
"raw_content": "# Example Domain\n\nThis domain is for use in documentation examples without needing permission. Avoid use in operations.\n\n[Learn more](https://iana.org/domains/example)",
"title": "Example Domain",
"url": "https://example.com"
}
]
}Explore more
| API | Price | Response |
|---|---|---|
Tavily URL Content ExtractionThis API Tavily | $0.012 per call | JSON |
Tavily | $0.012 per call | JSON |
Tavily | $0.012 per call | JSON |
Questions
Send a POST request to https://api.sandbase.ai/v1/api/tavily/extract with your SandBase API key in the Authorization header (Bearer) and the parameters as a JSON body. You can also try it in the Playground on this page without writing code.
$0.012 per call. Every organization gets 100 free calls per UTC day, shared across eligible SandBase APIs including this one; after that each call is billed at this price from your SandBase balance. There is no subscription.
Required: urls (array). Optional: extract_depth (one of basic, advanced) and include_images (boolean). The Parameters table and the Playground on this page list every field.
A JSON response returned in the same HTTP request (synchronous). The README on this page includes a sample response.
No. Tavily URL Content Extraction runs through SandBase, so one SandBase API key covers this endpoint and every other API and model in the catalog. You do not need to apply for Tavily API access separately.
Send urls as an array of public page addresses, not a single string or a search query. extract_depth accepts basic or advanced; include_images is a boolean. Review this endpoint's current schema instead of copying extra upstream options.
No. The current reference sample shows URL, title and image fields with failed_results, but does not demonstrate raw_content. This does not establish that page text is unavailable. Verify an approved response for your target before relying on a text field, full-page coverage or a downstream RAG workflow.
No. The documented envelope is outputs[0].data, and the reference describes results and failed_results. A completed run or HTTP success alone is not proof that every URL produced the content you wanted. Review the response and keep missing fields unknown rather than converting them to empty text.