PubMed scraper API: search results as JSON
Send any PubMed page URL with recipe=pubmed and get title, url as JSON. No selectors to write, no proxies to pick.
- 1 page type
- 2 fields
- 5 credits a page · $0.50 per 1,000
- Working · 100% in 7 days
5,000 free credits for 7 days, no card.
curl -G 'https://api.ujeebu.com/scrape' \
-H "ApiKey: $UJEEBU_API_KEY" \
--data-urlencode 'url=https://pubmed.ncbi.nlm.nih.gov/?term=large+language+models' \
--data-urlencode 'recipe=pubmed'
{
"result": {
"results": [
{
"title": "Large Language Models in Medicine: Applicatio...",
"url": "https://pubmed.ncbi.nlm.nih.gov/40520893/"
}
]
},
"success": true
}
What you get from PubMed
The fields of each page type and a real sample from our last validation, shortened.
Search and listing pages recipe=pubmed 5 credits a page
| Field | Type | Example |
|---|---|---|
results | list of 3+ items | |
results[].url | URL | https://pubmed.ncbi.nlm.nih.gov/40520893/ |
results[].title | text | Large Language Models in Medicine: Applications, Challenges, and Fu... |
Sample answer, validated 2026-11-16
{
"results": [
{
"url": "https://pubmed.ncbi.nlm.nih.gov/40520893/",
"title": "Large Language Models in Medicine: Applications, Challenges, and Future Directions."
},
{
"url": "https://pubmed.ncbi.nlm.nih.gov/38896652/",
"title": "Large Language Models in Orthopaedics: Definitions, Uses, and Limitations."
}
]
}
One call
Send the page URL and the recipe name to the Scrape API. The recipe brings the extract rules and the settings that work on PubMed (JavaScript rendering).
curl -G 'https://api.ujeebu.com/scrape' \
-H "ApiKey: $UJEEBU_API_KEY" \
--data-urlencode 'url=https://pubmed.ncbi.nlm.nih.gov/?term=large+language+models' \
--data-urlencode 'recipe=pubmed'
import os, requests
r = requests.get(
"https://api.ujeebu.com/scrape",
params={"url": "https://pubmed.ncbi.nlm.nih.gov/?term=large+language+models", "recipe": "pubmed"},
headers={"ApiKey": os.environ["UJEEBU_API_KEY"]},
timeout=120,
)
for item in r.json()["result"]["results"][:3]:
print(item["title"], item["url"])
# Example output:
# Large Language Models in Medicine: Ap... https://pubmed.ncbi.nlm.nih.gov/40520...
# Large Language Models in Orthopaedics... https://pubmed.ncbi.nlm.nih.gov/38896...
const params = new URLSearchParams({
url: "https://pubmed.ncbi.nlm.nih.gov/?term=large+language+models",
recipe: "pubmed",
});
const res = await fetch(`https://api.ujeebu.com/scrape?${params}`, {
headers: { ApiKey: process.env.UJEEBU_API_KEY },
});
const { result } = await res.json();
for (const item of result.results.slice(0, 3)) {
console.log(item.title, item.url);
}
// Example output:
// Large Language Models in Medicine: Ap... https://pubmed.ncbi.nlm.nih.gov/40520...
// Large Language Models in Orthopaedics... https://pubmed.ncbi.nlm.nih.gov/38896...
The answer is {"result": {...}} with the headers Ujb-Recipe and Ujb-Credits. Your own parameters always win over the recipe.
Search PubMed with parameters
One call, no URL: send recipe=pubmed/search with the parameters below. We build the PubMed search URL and answer the listing JSON, at the same credits as the page call.
| Parameter | Type | Values |
|---|---|---|
query required | string | Search words, as typed in the PubMed search box. |
One call (query=solid-state batteries)
curl -G 'https://api.ujeebu.com/scrape' \
-H 'ApiKey: YOUR_API_KEY' \
-d 'recipe=pubmed/search' \
--data-urlencode 'query=solid-state batteries'
import os, requests
r = requests.get(
"https://api.ujeebu.com/scrape",
params={"recipe": "pubmed/search", "query": "solid-state batteries"},
headers={"ApiKey": os.environ["UJEEBU_API_KEY"]},
timeout=120,
)
for item in r.json()["result"]["results"][:3]:
print(item["title"], item["url"])
# Example output:
# Large Language Models in Medicine: Ap... https://pubmed.ncbi.nlm.nih.gov/40520...
# Large Language Models in Orthopaedics... https://pubmed.ncbi.nlm.nih.gov/38896...
const params = new URLSearchParams({
recipe: "pubmed/search",
query: "solid-state batteries",
});
const res = await fetch(`https://api.ujeebu.com/scrape?${params}`, {
headers: { ApiKey: process.env.UJEEBU_API_KEY },
});
const { result } = await res.json();
for (const item of result.results.slice(0, 3)) {
console.log(item.title, item.url);
}
// Example output:
// Large Language Models in Medicine: Ap... https://pubmed.ncbi.nlm.nih.gov/40520...
// Large Language Models in Orthopaedics... https://pubmed.ncbi.nlm.nih.gov/38896...
The answer carries the built URL in the Ujb-Url header; a missing or invalid parameter answers 400 naming it (free). Have the URL already? Send it with recipe=pubmed/listing. Search with parameters in the docs
What people build with PubMed data
Research datasets
Build datasets of public pages for analysis, search or LLM pipelines.
Learn more
Price
Same price as the page call, no fee for the recipe. Only successful responses are billed; blocked, timed-out and failed requests are free.
| Page type | Credits a page | Per 1,000, Starter | Per 1,000, Pro |
|---|---|---|---|
| Search results | 5 | $0.82 | $0.50 |
FAQ
What does the PubMed recipe return?
JSON with 2 fields from search results (recipe=pubmed), for example title, url. The answer is {"result": {...}} with the fields shown on this page.
How much does it cost to scrape PubMed?
5 credits a page ($0.50 per 1,000 on Pro, $0.82 on Starter). Same price as the page call, no fee for the recipe. Only successful responses are billed; blocked, timed-out and failed requests are free.
Do I need proxies or a headless browser?
No. The recipe brings the settings that work on PubMed: JavaScript rendering. Send only the URL and recipe=pubmed.
What happens when PubMed changes its pages?
We check PubMed every day and re-validate the recipe on real pages every week, then fix it when the site changes. Until a fix is live, the API answers 404 "the recipe has no working version" instead of empty data, and that call is free.
Can I search PubMed with parameters?
Yes, in one call: send recipe=pubmed/search with query and no URL. We build the PubMed search URL (returned in the Ujb-Url header), apply recipe=pubmed/listing and answer its JSON, at the same credits as the page call. A missing or invalid parameter answers 400 naming it, and that call is free.
Related recipes
- MDPISearch results · 5 credits
LinkedInSearch results, Detail page, Profile · 5 credits
AmazonSearch results · 5 credits
X (Twitter)Profile · 5 credits
InstagramProfile · 5 credits
TikTokProfile · 5 credits
Logos and names belong to their owners. Ujeebu is not affiliated with PubMed.
Get PubMed data as JSON today.
5,000 free credits for 7 days. No card.