Presearch AI Results Scraper
Under maintenancePricing
from $1.99 / 1,000 search results
Presearch AI Results Scraper
Under maintenanceScrapes Presearch PandaChat answers with cited sources, organic search context, related questions, and normalized metadata.
Pricing
from $1.99 / 1,000 search results
Rating
0.0
(0)
Developer
Search API
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
15 days ago
Last modified
Categories
Share
This Actor submits one or more prompts to Presearch and extracts the PandaChat answer together with the public organic results that support it.
Input
prompt(required): the primary search prompt.prompts: optional additional prompts. Duplicates are removed.maxItems: maximum number of prompts to process, from 1 to 100.maxRequestRetries: bounded retries per prompt, from 0 to 5.proxyConfiguration: Apify Proxy or custom proxy configuration.
Dataset records
Each successful prompt produces one normalized record containing a stable ID, prompt, answer text and available HTML, answer and word counts, citation indexes, cited source objects, all organic source objects, related questions, the Presearch result ID, provider, search URL, and scrape timestamp.
Each source can include its position, title, URL, domain, snippet, provider, and favicon URL. Unavailable fields are omitted rather than filled with placeholders.
The Actor listens for Presearch's public JSON and event-stream responses before navigation. It validates response status and content type, maps only documented dataset fields, and uses the rendered PandaChat card as a fallback when the summary response is unavailable. It never stores raw API payloads.
Presearch currently applies Cloudflare access controls to many datacenter and workstation IPs. Direct runs detect that response and fail closed. Use an authorized Residential proxy when permitted; the Actor does not attempt to solve or bypass interactive challenges. Proxy credentials are never read from environment variables or written to output.
If Presearch does not return a usable PandaChat answer after bounded retries, the run fails instead of publishing a citation-only or fabricated record.
Local validation
