web-search

The web-search skill provides AI-powered web search and content extraction through the inference.sh CLI, integrating the Tavily and Exa applications. It allows users to perform intelligent web searches, extract relevant content from multiple URLs, and retrieve direct answers to factual queries. By automating these processes, the skill addresses the challenge of efficiently gathering and processing information from the internet, which is especially useful for research, data aggregation, and building AI agents capable of knowledge retrieval.

This skill offers several core features: AI-enhanced search via Tavily and Exa apps, content extraction from single or multiple web pages, direct factual answering, and integration into research workflows with large language models (LLMs). Users can interact with the skill through simple CLI commands to execute searches, extract data, and analyze results. It supports workflows such as search plus LLM analysis or extraction plus summarization, making it adaptable to research pipelines, retrieval-augmented generation (RAG) tasks, and fact-checking. The skill also provides structured outputs, including clean text, images, and source links, enabling easy downstream processing.

The web-search skill is designed for researchers, AI developers, content aggregators, and anyone needing reliable and structured web information. Common use cases include conducting in-depth research, verifying claims, gathering content for RAG systems, and feeding data into AI agents. It targets users who require precise, AI-assisted web search and extraction capabilities while maintaining flexibility for integration with other tools and models within the inference.sh ecosystem. Its CLI-based design ensures compatibility across systems where Bash is available, making it suitable for both individual researchers and larger automated pipelines.