by darko893 · v0.1.0
Extract structured data or clean text from public web pages with a plain-language prompt. Unreadable pages return an honest error code instead of invented content, and failed reads are not charged.
This community listing does not yet include every recommended support, privacy, pricing, and permission disclosure. Review the available package permissions before installing.
Available inside your emploidai workspace after installation.
Available inside your emploidai workspace after installation.
Extract structured data or clean text from public web pages inside your Dify apps and agents, using Haunt.
Give the tool a URL and a plain-language prompt ("the product name, price and stock status") and it returns structured JSON. No CSS selectors, no parsing rules.
The part that matters for agents: honest failure. When a page cannot be read (bot wall, login wall, missing page), the tool returns a plain error code such as access_denied, login_required, or not_found instead of invented content, so your workflow can branch on the failure instead of acting on made-up data. Failed reads are not charged.
Add the Haunt Extract tool to an agent or workflow. Parameters:
Success returns JSON like:
{"stories": [{"title": "...", "points": 572}]}
An unreadable page returns text like:
Haunt could not read https://example.com/admin. error_code=login_required. This page needs a login. This read was not charged.
hauntapi.com only.https://github.com/Darko893/dify-haunt
support@hauntapi.com or https://hauntapi.com/docs?utm_source=dify&utm_medium=integration&utm_campaign=sweep-2026-07