This project provides a dual-engine web search skill for AI agents like OpenClaw/Pi, integrating Grok AI and Tavily search with FireCrawl as a fallback. It aims to enhance search accuracy through cross-validation and offers a flexible, pure Shell implementation, benefiting AI developers and power users seeking robust web data retrieval.

How It Works

The skill employs a suite of shell scripts (grok-search.sh, tavily-search.sh, web-fetch.sh, dual-search.sh) to manage web searches. It prioritizes Grok AI via OpenAI Compatible APIs for deep searching and uses Tavily for structured results, running them in parallel via dual-search.sh for cross-validation. A three-tier fallback mechanism in web-fetch.sh (Tavily Extract -> FireCrawl Scrape) ensures content retrieval even if primary methods fail. The TavilyProxyManager supports aggregating multiple Tavily API keys for load balancing and redundancy.

Quick Start & Requirements

• Primary install: Clone the repository, navigate to the directory, and copy the environment template:

  git clone https://github.com/ckckck/UltimateSearchSkill.git
  cd UltimateSearchSkill
  cp .env.example .env
  

• Prerequisites: Docker and Docker Compose, curl, jq. API credentials for OpenAI Compatible API (for Grok), Tavily, and optionally FireCrawl are required.
• Setup: Configure API keys and other settings in the .env file.
• Links: Tavily (https://www.tavily.com/), FireCrawl (https://www.firecrawl.dev/).

Highlighted Details

• Dual-Engine Search: Leverages Grok AI and Tavily concurrently for cross-validated search results.
• Multi-Account Aggregation: Tavily API keys can be aggregated using TavilyProxyManager.
• Graceful Degradation: web-fetch.sh automatically falls back from Tavily Extract to FireCrawl Scrape.
• Pure Shell Implementation: Relies on Bash scripting and Docker, minimizing external dependencies.
• Security: Binds ports to 127.0.0.1, includes API authentication, and supports SSH tunneling for management.
• Search Methodology: Incorporates GrokSearch MCP's search planning framework and evidence standards.

Maintenance & Community

The README does not detail specific contributors, sponsorships, or community channels like Discord/Slack.

Licensing & Compatibility

The project is released under the MIT License, which permits commercial use and integration into closed-source projects.

Limitations & Caveats

The grok2api method is designated as legacy/experimental, with a strong recommendation to use Grok via OpenAI Compatible APIs. Integration requires specific setup steps within the OpenClaw/Pi agent framework, including modifying PATH and agent configurations. The effectiveness of Grok search depends on the availability and configuration of compatible OpenAI API endpoints.