Yes, several SERP APIs return the raw HTML alongside the parsed results: of the seven checked here, six return it with the parsed JSON, as a string or as a link, and DataForSEO serves it from a separate call on the same task.
- Inline or a link. Bright Data, Oxylabs, Zyte and ScrapingBee put the HTML in the JSON as a string. SerpApi and cloro return a URL to a stored HTML file. DataForSEO returns it from its own
task_get/htmlendpoint. - Two vendors state the price of the HTML. cloro says HTML is included at no extra cost, and DataForSEO charges only for posting the task. The other five docs pages do not mention a separate charge.
- Retention ranges from 24 hours to 31 days. cloro's HTML URLs expire after 24 hours, DataForSEO results stay available for 7 days, and SerpApi's Search Archive API returns a search for up to 31 days. Inline HTML is kept only as long as you keep it.
Does any SERP API return the raw HTML alongside the parsed results?
Each row comes from the vendor's own documentation, read on 16 November 2026. "Not stated" means the page we opened does not say.
| API | How to get the HTML | In the same response as the parsed JSON? | Extra cost | Retention |
|---|---|---|---|---|
| Bright Data | brd_json=html or header x-unblock-data-format: html | Yes, "full JSON with a structure holding the full HTML as string" | Not stated | Inline |
| Oxylabs | output: ["html", "json"] on /v1/scrape/google/search | Yes, output is a list of formats | Not stated | Inline |
| Zyte | include: ["html", "organic"] on /v1/search | Yes, html beside organicResults | Not stated | Inline |
| ScrapingBee | add_html=true | Yes, in a full_html field | Not stated | Inline |
| SerpApi | search_metadata.raw_html_file, or output=html | Yes, as a URL. output=html returns HTML instead of JSON | Not stated | Not stated for the file; a search stays in the Search Archive API for 31 days |
| cloro | include.html: true | Yes, result.html is an array of CDN URLs, one per results page | None | 24 hours |
| DataForSEO | GET /v3/serp/google/organic/task_get/html/$id | No, a separate call on the same task | None, "charged only for posting a task" | 7 days |
Serper is not in the table. Its public pages show JSON responses and document no HTML option that we could find.
How do the seven APIs differ?
Inline HTML is the simplest to handle: one response, one parse, nothing to fetch later. Bright Data documents brd_json=html for Google web search as the full parsed JSON with the HTML inside it as a string. Zyte's search endpoint takes an include list. "html" returns the rendered page and "organic" returns parsed organicResults, and its default is ["html"] alone. Oxylabs' Web API takes output as a list of markdown, html, json and screenshot, with ["html"] as the default. ScrapingBee's add_html defaults to false, and its docs say it is "useful if you want to get some information we don't return in the JSON response". The cost of inline HTML is size: every response carries a full results page.
A link to the HTML keeps the JSON small. SerpApi's example responses carry search_metadata.raw_html_file, a serpapi.com/searches/...html URL, next to json_endpoint. Its output parameter is a switch: json (the default), html "to get the raw HTML retrieved" or md. With output=html you get the page and no parsed fields in that response.
cloro's Google Search endpoint returns result.html as an array when include.html is true, one CDN URL for each page scraped, so a 3-page request gives 3 URLs. The output formats guide says the URL expires after 24 hours, so download the file if you need it later. The providers guide says "HTML responses are included at no additional cost".
A separate call suits a queue. DataForSEO's flow posts a task and reads the result later. Its tasks_ready endpoint gives one task three result URLs, endpoint_regular, endpoint_advanced and endpoint_html, and its docs say you "can get the results of the task within the next 7 days for free".
Why would you want the raw HTML?
Three reasons come up in the vendors' own docs.
- Proof of what the page showed. The output formats guide describes the HTML as "proof of what the AI provider actually displayed", which a client can check. A rank report can link each position to the page it came from.
- A block the API does not parse. SerpApi says its HTML output supports "features not yet supported by SerpApi", and ScrapingBee says nearly the same. With the HTML saved, you can parse a new block the day you need it, and parse old pages again later.
- Debugging the parser. SerpApi says HTML output "is useful to debug JSON results". When a field is empty, the HTML shows whether the block was missing from the page or the parser missed it.
If you only need the HTML now and then, a link or a separate call costs less bandwidth than inline HTML on every response. If you store every page for an audit trail, inline HTML saves a second request, and a link with a short expiry needs a download job.
How do you fetch the HTML with the parsed JSON in Python?
This request asks the Google Search endpoint above for two pages with HTML. It prints the parsed organic results and saves each page's HTML before the URLs expire. It uses requests and reads the API key from an environment variable.
import os import requests resp = requests.post( "https://api.cloro.dev/v1/monitor/google", headers={"Authorization": f"Bearer {os.environ['CLORO_API_KEY']}"}, json={ "query": "best laptops for programming", "country": "US", "pages": 2, "include": {"html": True}, }, timeout=180, ) resp.raise_for_status() result = resp.json()["result"] # Parsed results from both pages, in one list for r in result["organicResults"]: print(r["page"], r["position"], r["title"], r["link"]) # One CDN URL per results page; each expires after 24 hours for page_number, url in enumerate(result["html"], start=1): html = requests.get(url, timeout=60).text with open(f"serp-page-{page_number}.html", "w", encoding="utf-8") as f: f.write(html) print(f"page {page_number}: {len(html):,} characters of HTML saved") This sync request costs 7 credits: 3 for the search, 2 for the second page and 2 for the sync surcharge. The HTML adds nothing. On the async API (POST /v1/async/task) the same search costs 5 credits.
The saved HTML includes scripts and page chrome. To read a block the JSON does not have, load the file into any HTML parser, such as BeautifulSoup, and select the block. Open the file with JavaScript off, because the page's own scripts can break it in a browser.
Which one should you pick?
Pick by how you use the HTML:
- Every response, stored for an audit: an inline option (Bright Data, Oxylabs, Zyte, ScrapingBee) saves a second request.
- Proof on demand with small JSON: a link (SerpApi's
raw_html_file, or cloro'sresult.html, which you download within 24 hours). - Batch jobs read later: DataForSEO's task flow, with 7 days to collect the HTML.
Check the cost line for your vendor. Only two of the seven docs pages state that the HTML is free, and every vendor charges for the search itself.
How was this checked?
We opened each vendor's documentation page linked in the table on 16 November 2026 and copied the parameter names and the wording on cost and retention. A "Not stated" cell means the page did not say either way. Vendor docs change, so check them before you build on a field.
This post is published by cloro, which sells one of the APIs in the table. The row for its own API comes from its own docs.
