free-search-aggregator

name	free-search-aggregator
description	Search across multiple web providers with automatic failover. Use when the user needs quota-aware, multi-provider web search.
license	MIT
compatibility	claude-code opencode github-copilot devin pi cursor

Free Search Aggregator

Reliable web search with provider failover and quota awareness.

Providers

Default order:

brave -> exa -> tavily -> duckduckgo -> bing_html -> mojeek -> serper -> searchapi -> google_cse -> baidu -> wikipedia

No-key providers: DuckDuckGo, Bing RSS, Mojeek, Wikipedia.

Commands

# Normal search
scripts/search "latest AI agent frameworks" --max-results 5

# Task search (multi-query)
scripts/search task "@dual Compare Claude vs GPT-4"

# Deep research
scripts/search task "@deep autonomous vehicle safety" --max-results 8 --max-queries 10

# Quota and health
scripts/status
scripts/remaining --real
python3 -m free_search health
python3 -m free_search doctor

Rules

• Default mode uses one worker for cost control.
• Use @dual or @deep only for research tasks.
• Results stored in memory/search-cache/ and memory/search-reports/.

Output

• Search: query, provider, results, attempted providers, quota
• Task search: task, queries, grouped and merged results