ChatGPT + Hydrafetch
Let ChatGPT work from live pages.
Add the Hydrafetch app and ChatGPT can read any page in full, search the web and read what it finds, pull recent news and extract data into a table. Each result carries its address, so answers come with their sources.
ChatGPT
with Hydrafetch
[ 01 / Set up ]
Open the app, then connect.
The app is the quick route. The custom connector does the same thing if you prefer to add it by address.
From the ChatGPT apps list
- 1
Open Hydrafetch in ChatGPT's apps and choose Connect.
Open in ChatGPT - 2
Sign in to Hydrafetch and approve.
Add it as a custom connector instead (paid ChatGPT plans)
- 1
Open Settings, then Apps and Connectors, then Advanced settings, and turn on developer mode.
- 2
Create a connector, name it Hydrafetch and paste this URL.
Paste thishttps://api.hydrafetch.com/mcp - 3
Choose OAuth for authentication, create it, then sign in and approve.
[ 02 / What to ask ]
Ask in plain words. ChatGPT picks the tool.
These are the kinds of request that use each tool, with what the call costs.
Read this article and tell me what it says about pricing.
scrape1 creditFind the best explanations of connection pool sizing and compare them.
search1 credit plus 1 per scraped resultWhat is the news on battery storage this week?
news1 credit plus 1 per article readList every page on this site so I can see what it covers.
map1 creditTurn the plans on these pricing pages into one comparison table.
extract5 credits per URLTake a screenshot of this page.
screenshot5 credits
[ 03 / What comes back ]
Five pages read, from one search.
This is what the search tool returns for one question: the ranked results, each already read and turned into clean text, with a confidence figure for each page.
PostgreSQL Connection Poolingcompilenrun.com1,745 wordsconfidence 0.98
PostgreSQL Connection Pooling with PgBouncer: A Complete Guidedev.to1,099 wordsconfidence 0.99
Connection pooling strategy using PgBouncer in Azure Databaselearn.microsoft.com2,734 wordsconfidence 1
Pooling Best Practices, PostgreSQL SME Cookbookpostgresql.codeguides.io351 wordsconfidence 0.66
PostgreSQL Connection Pool Sizing: A Practical Guide for Web Appsarmordb.org1,964 wordsconfidence 0.99
the result of the search tool, captured from the live API · every page arrives as clean text with its address
[ 04 / Questions ]
What people ask about ChatGPT and Hydrafetch.
Add the Hydrafetch app in ChatGPT and choose Connect. ChatGPT can then read a page you name, search the web and read the results, and work from the text of those pages.
Yes. Open Hydrafetch in the ChatGPT apps list, choose Connect, then sign in to Hydrafetch and approve. On paid ChatGPT plans you can also add it as a custom connector with its address.
The same credits as the API. Reading a page is 1 credit, a search is 1 credit plus 1 per scraped result, and a call that fails costs nothing. Your ChatGPT plan is separate.
Read a specific page in full as clean text, crawl or map a whole site, extract fields into a table from several pages, search news by date, search open jobs with pay, and return brand data for a domain.
No. You sign in to Hydrafetch in the browser and approve the connection. The connection appears in your Hydrafetch dashboard, where you can remove it.
[ 05 / Keep going ]
Other ways to connect.
[ Start ]
Clean web data is one call away.
500 free credits, no card required. Failures are never billed.
Success rate
Median scrape
ms