BrightData Web Fetcher — TypingMind Plugin

Automatically search or fetch web content using BrightData APIs (SERP & Web Unlocker) based on natural language instructions and parse large HTML or structured responses through OpenAI to extract relevant, clean JSON data.

Features

• Parallel Web Search & Fetch
  Search Google using BrightData SERP API (multiple queries in parallel), or fetch webpage content with Web Unlocker API (multiple URLs in parallel).

• Full Google gl and hl RAG Support
  Uses all Google-supported gl (country) and hl (language) codes for precise localization and retrieval-augmented generation (RAG). The AI is aware it can use any supported value.

• Content Extraction for All Web Content
  For HTML, uses OpenAI to extract only the main readable content as structured JSON (works for articles, docs, blogs, forums, etc). For JSON, XML, JS, CSS, or plain text, returns the raw content as-is.

• Minimal Commands
  Plain, natural language instructions (“find Paris climate news”), handled contextually without the need for explicit commands.

• Advanced Pagination & Filtering
  Supports start, num, tbm (news/shopping), and job search with ibp. Combine as needed for advanced queries.

• Highly Efficient for Research
  Parallelizes OpenAI extraction calls for fast processing of many URLs or searches.

• Secure & Modular
  Separate SERP and Web Unlocker keys/zones, and secure OpenAI key management.

OpenAI

• Why Use OpenAI for Extraction?
  Large web pages and search result HTML can exceed model context windows. This plugin parses and extracts only the structured, important content via OpenAI before handing results to the chat, minimizing context/token usage.

• Best Models
  Defaults to high-context models (like gpt-4.1-mini) for reliability with large or complex results.

Setup

1. In TypingMind, go to Plugins → Import plugins → GitHub URL, and paste the repo URL (or upload plugin scripts).
2. Open plugin Settings and provide:
   • BrightData SERP API Key
   • SERP Zone ID
   • BrightData Web Unlocker API Key
   • Web Unlocker Zone ID
   • OpenAI API Key
   • OpenAI Model to use
3. Start your research! Ask natural-language questions—the plugin handles all search/fetch actions and delivers clean, readable, local-language-aware results.

Requirements

• BrightData account with SERP & Web Unlocker access and sufficient credit.
• Valid BrightData API keys and zone IDs.
• Valid OpenAI API Key and sufficient credit.
• TypingMind platform with plugin support.

Supported Parameters

• action: "search" (Google SERP API) or "fetch" (Web Unlocker API)
• query: Single query OR multiple queries (array) for searching in parallel
• url: Single URL OR multiple URLs (array) for fetching in parallel
• tbm: "nws" (news), "shop" (shopping)
• ibp: "htl;jobs" for jobs
• gl: Any valid Google country code
• hl: Any valid Google language code
• start: Results offset (pagination)
• num: Number of results per query

The RAG (retrieval-augmented generation) coverage includes all Google-supported gl and hl codes.

---