SearchWhisperer 2.6.1

AI-powered search query optimization in any language

v2.6.1 Multi-language | Expert terminology | Real-time streaming

Support Independent Research Tools

Which language is your search in?

We detected multiple possible languages for your query.

Or specify your language:

Loading...

Analyzing your query...

About

Search Whisperer helps you construct effective Google queries for better search results.

Version History

v2.6.1 Current

Full UI localization - Tabs, personas, and descriptions translate to your query's language
Specificity Intelligence - Detailed queries get deeper suggestions that preserve your context (location, time, demographic)
Keyword explanations - Each suggestion now shows WHY it helps (e.g., "Broader search", "Scientific term")
Multi-language support - Search in any language: German, Dutch, French, Japanese, and more
Automatic language detection - AI detects your query's language with confidence scoring
Smart clarification - When language is ambiguous, a dialog asks you to confirm
Native expert terminology - Keywords generated in your language (e.g., "Demenz Ätiologie" for German medical queries)
Localized explanations - Google dork explanations written in your query's language
Mixed English technical terms where universally used (MRI, DNA, API)

v2.4

75% faster Full Repair - Reduced from 11-12s to 3-4s with optimized prompts
3-5x faster Hotfix - Switched to Haiku model for rapid response
Fixed streaming reliability - Full Repair now completes without fallback delays
Inline suggestion editing - Click pencil icon to edit any suggestion before using
Google dork validation - Fixed wildcard operators (no more invalid truncation)
Removed content filtering - All queries analyzed objectively without refusals
Streamlined prompts across all LLM services for speed

v2.3

Expert terminology intelligence - Now predicts real jargon from research papers and technical docs
Domain-aware keywords - Identifies field and generates 8-12 expert-level terms
Includes acronyms, tools, frameworks (PyTorch, BERT, fMRI, etc.)
Uses 2024-2025 terminology (LLMs, RAG, prompt engineering)
Mixes concepts, methods, tools, and metrics for comprehensive coverage
Prompts enhanced across Claude, OpenAI, and Groq services

v2.2

67% faster Full Repair - Reduced from 14.8s to 4.9s total time
Real-time streaming - Results appear as generated (JSONL format)
Parallel generation - Two simultaneous API calls for 2x throughput
Switched Full Repair to Haiku for 3-5x speed boost
Non-blocking query analysis (starts streaming instantly)
Reduced token usage (1500 → 1000) for faster generation

v2.1

50% faster query analysis (5.4s → 2.6s)
Upgraded to Claude 4.5 Haiku for first-round analysis
Optimized token usage (800 → 300 tokens)
Updated to latest Claude API models

v2.0

Complete UI redesign with enhanced tabs
Hotfix and Full Repair modes for query optimization
Library tab for additional resources
Performance improvements and skeleton loaders

© 2025 Search Whisperer (Henk van Ess)