All Jina AI APIs (Reader, Search, Embed, Rerank, and more) as powerful Unix CLI commands. Ideal for web scraping, semantic search, and RAG (Retrieval-Augmented Generation) workflows.
Capabilities
jina search <QUERY>: Performs deep web searches. Note: If this command fails with auth errors, use the curl alternative below.jina read <URL>: Extracts clean, markdown-ready content from any webpage.jina embed <TEXT>: Generates text embeddings for semantic similarity or storage.jina rerank <QUERY>: Re-ranks documents based on relevance to a search query.jina screenshot <URL>: Captures high-quality screenshots of webpages.jina pdf <URL>: Extracts structured data (tables, charts, formulas) from PDF documents.jina expand <QUERY>: Expand queries for search.jina bibtex <QUERY>: Search bibtex citations.jina primer: Get system context (time, location, authenticated user).
Usage Examples
Web Search (Fallback if jina search fails)
curl -H "Authorization: Bearer $JINA_API_KEY" "https://s.jina.ai/<QUERY>"
Web Reading for LLMs
jina read https://jina.ai/reader | pbcopy
Advanced Search & Reranking (RAG)
jina search "latest trends in AI agents" --json | jina rerank "autonomous systems" --top-n 3
Academic Research
jina search "transformer architectures" --arxiv --time 2024
jina bibtex "Attention is All You Need"
Security & Configuration
- API Token: The
JINA_API_KEYis pre-configured in~/.zshrc. - Environment: Ensure the command is executed in an environment that has access to the shell profile.