BrowseAI-Dev

Overview

BrowseAI Dev is a research infrastructure for AI agents that provides web search with verification capabilities. It implements a comprehensive verification pipeline that extracts claims from web pages and verifies them against actual source text. The system provides confidence scores, citations, and anti-hallucination features. It offers different depth modes for research thoroughness and supports streaming APIs for real-time feedback. The MCP server exposes these capabilities through the Model Context Protocol, allowing AI agents to access reliable research information.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• search

  Basic web search functionality

  Try:Search for information about quantum computingTry:Find recent news about artificial intelligence
• answer

  Verified search with citations and confidence scores

  Try:Answer what is quantum computing with sourcesTry:Find information about climate change with verified claims
• extract

  Extract specific information from web pages

  Try:Extract the main findings from this research paperTry:Get the contact information from this company website
• compare

  Compare multiple sources or options

  Try:Compare different approaches to renewable energyTry:Contrast the features of these two smartphone models
• clarity

  Anti-hallucination verification tool

  Try:Verify these claims about AI technologyTry:Check if these facts about space exploration are accurate
• session

  Create a persistent research session

  Try:Start a research session about machine learningTry:Create a session to track information about electric vehicles
• feedback

  Submit feedback on search results

  Try:Mark this result as goodTry:Flag this claim as incorrect

Comparable tools

Installation

npm install browseai-dev

pip install browseaidev

pip install langchain-browseaidev

{
  "mcpServers": {
    "browseai-dev": {
      "command": "npx",
      "args": ["-y", "@browseai-dev/mcp-server"]
    }
  }
}

FAQ

How does the verification system work?

Claims are extracted from web pages and verified against actual source text through atomic claim decomposition, hybrid retrieval, semantic reranking, and multi-source consensus. Each claim receives a confidence score based on verification strength.

What are the depth modes?

Three levels: 'fast' (single pass), 'thorough' (multiple verification passes with iterative queries), and 'deep' (multi-step agentic research with gap analysis and targeted follow-ups).

Compare BrowseAI-Dev with