{
  "id": 3027387,
  "title": "Give Your AI Agent a Scientist's Library. a Science MCP Server",
  "url": "https://urgent.news/2026/08/24/give-your-ai-agent-a-scientists-library-a-science-mcp-server",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-24T13:29:29.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/valyuai/give-your-ai-agent-a-scientists-library-a-science-mcp-server-4pdb"
  },
  "original_language": "en",
  "account": "Most research agent demos search abstracts and call it a literature review. The abstract only reveals a Phase 3 melanoma immunotherapy trial's endpoint, not the imaging protocol or supplementary data. This guide wires a hosted science MCP server into an AI agent, scopes it to real scientific collections, and runs a controlled before-and-after test where only one variable changes. Valyu is a search API designed for AI agents, providing full text and structured metadata across biomedical literature, clinical trial registries, patents, financial filings, and the open web with resolvable identifiers. The hosted MCP server is a plain HTTP endpoint with no local process, Node.js, or mcp-remote required. The MCP tools only need a query and max_num_results, with no source filtering. Scope results using the included_sources parameter in the API or SDK, then check result.source on every hit. Include_abstracts=True expands PubMed to the full abstract corpus, while the default restricts to papers with available full text. Every claim receives a resolvable identifier, such as a DOI. The API returns result.doi - read it, don't prompt for it. To use Valyu, obtain an API key from platform.valyu.ai, which offers $10 in free credits or $20 with a work email. Connect through Claude Desktop, Cursor, or any MCP client with Python 3 and pip install valyu. The hosted server is remote HTTP, connecting via https://mcp.valyu.ai/mcp?valyuApiKey=YOUR_API_KEY. The server exposes 11 tools: valyu_search, valyu_academic_search, valyu_bio_search, valyu_patents, valyu_contents, valyu_datasources, valyu_financial_search, valyu_sec_search, valyu_company_research, valyu_economics_search, and valyu_datasources_categories. Scoping sources can be done through presets (curated bundles) or dataset IDs (individual collections). Presets include academic, finance, patent, health, genomics, chemistry, physics, legal, politics, transportation, pulse, cybersecurity, environment, automotive, compliance, and medical. Watch preset boundaries, as academic covers literature and preprints only, while health includes clinical trial registry records. The before-and-after test compares full-text-only retrieval against broad abstract-level retrieval, not the same search with and without full text.",
  "summary": "Most \"research agent\" demos are searching abstracts and calling it literature review. The abstract tells you a Phase 3 melanoma immunotherapy trial hit its endpoint. It does not tell you the imaging protocol used to assess tumour response that lives in the methods section, or a supplementary table, or a figure caption. This walks through wiring a hosted science MCP server into your AI agent,…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}