Search CLI from Paperfoot is a command-line search utility for AI agents and the Model Context Protocol. It aggregates results from Brave, Serper, and Exa, runs parallel queries, and returns structured JSON tailored for agent consumption, enabling fast retrieval of web pages, academic papers, and social posts. Notable features include 14 specialized modes, semantic exit codes, parallel execution across providers. Target users are AI developers and researchers who need a consistent MCP-native retrieval tool for agent workflows.

What tasks can you actually use it for?

The tool functions as a unified query layer for agent-driven web retrieval. Search accepts a prompt-style query and returns aggregated hits from Brave, Serper, and Exa, covering general web lookups, news, academic research, and social media. Fourteen specialized modes let agents target the desired source mix for a task. Output is emitted as structured JSON, which agents can parse directly to extract:

• titles
• snippets
• URLs
• provider metadata

How accurate are the outputs compared to doing it manually?

Search aggregates provider responses very quickly, typically returning combined results in under two seconds thanks to parallel querying. The tool uses semantic exit codes to indicate empty or error conditions, which helps agent decision logic. Accuracy of any individual result depends on the underlying provider; the tool relays provider content rather than synthesizing claims. Teams that need verified facts should follow returned URLs or snippets for independent validation before using results in critical automation.

Does it require technical knowledge to get useful results?

Search expects a technical setup: installation paths include Homebrew, Cargo, or a shell installer, and it runs on systems that support MCP, including macOS, Linux, and Windows. Users must supply API keys for Brave, Serper, or Exa via environment variables or a configuration file. Integrators configure the CLI as an MCP tool for agent platforms such as Claude Desktop, so developers and power users can embed the tool into automated retrieval pipelines.

A practical, community-backed tool for embedding web retrieval into agents

Paperfoot's Search is a practical option for teams embedding web retrieval into agent pipelines, supported by broad adoption and frequent citations within the MCP developer community. Expect an operational tool that integrates into existing agent logic rather than a substitute for independent research. For reliable outputs, include explicit verification steps in your pipeline that follow up returned results with original source checks before using them in decision-making.