Bgpt.pro
FreemiumMCP server for searching scientific papers with structured experimental data from full-text studies. Returns 25+ fields per paper. Works with Claude, Cursor, Cline. 50 free searches, then $0.01/result.
What does this tool do?
BGPT MCP is a specialized API that integrates scientific paper search directly into AI coding tools like Claude, Cursor, and Cline. Rather than returning paper summaries or abstracts, it extracts granular experimental data from full-text studies—including methods, sample sizes, results, limitations, and funding information. The tool operates as an MCP (Model Context Protocol) server, meaning your AI assistant can autonomously search for and retrieve structured paper data with 25+ fields per result. This is designed for researchers, developers, and data scientists who need evidence-grounded answers in real-time without leaving their IDE or chat interface.
AI analysis from Feb 23, 2026
Key Features
- MCP server integration compatible with Claude Desktop, Cursor, Cline, and any MCP-compatible AI client
- Full-text paper parsing with extraction of 25+ structured fields including methods, results, limitations, and quality scores
- Search parameters including query strings, result count (1-100), and time filtering (days_back parameter)
- Critical evaluation metadata per paper: falsifiability assessment, study blindspots, bias detection, and conflict of interest flags
- Technical metadata: experimental models, software tools used, data availability statements, code/data links, lab names, and funding sources
- SSE (Server-Sent Events) streaming endpoint for real-time result delivery
- Free tier tracking and automatic paid upgrade prompts with Stripe integration for one-click billing
Use Cases
- 1Researchers validating experimental methodologies by searching for prior studies with comparable sample sizes and techniques
- 2Biotech developers finding CRISPR, gene therapy, or drug delivery papers with actual experimental results and reproducibility assessments
- 3PhD students conducting literature reviews while working in their IDE, pulling structured data on study limitations and conflicts of interest
- 4Data scientists building models that require grounding in peer-reviewed experimental data rather than general summaries
- 5Medical professionals evaluating clinical trials and real-world study outcomes during diagnostic or treatment planning conversations
- 6Grant writers gathering evidence of feasibility and novelty by accessing extracted study context and funding sources from published papers
Pros & Cons
Advantages
- Extraction of raw experimental data (not summaries) directly from full-text papers, reducing hallucination and increasing evidence quality in AI responses
- Generous free tier with 50 searches before any payment required, lowering barrier to entry for casual or academic users
- Transparent, granular billing at $0.01 per result returned (not per search), aligning cost with actual usage rather than speculation
- Seamless integration into existing workflows—configured in 30 seconds into Claude Desktop, Cursor, or CLI-based tools without additional authentication friction
- Rich metadata per paper (25+ fields including study blindspots, falsifiability, quality scores, and code/data links) enabling critical evaluation, not just retrieval
Limitations
- Limited to scientific papers only—no support for preprints, unpublished datasets, or grey literature, excluding emerging research and non-peer-reviewed work
- Unclear data freshness and database scope; no information on how many papers are indexed, publication date range, or update frequency
- Pricing scales poorly for high-volume users; a researcher running 1,000 searches monthly at 10 results each pays $100, with no bulk discounts or enterprise tier mentioned
- Dependency on MCP protocol limits adoption—users without Claude, Cursor, Cline, or compatible tools cannot access the service
- No API documentation publicly visible on the marketing page; developers must reverse-engineer parameter behavior or contact support for details
Pricing Details
Free tier: 50 searches included, no API key required, full access to all 25+ metadata fields. Paid tier: $0.01 per result returned (not per search), billed daily by Stripe. Billing is granular—if a search returns 7 results, you pay $0.07; if zero results, no charge. No monthly minimums, subscription caps, or enterprise pricing mentioned. Cancel anytime through Stripe customer portal.
Who is this for?
Research scientists, biotech/pharma developers, PhD students conducting systematic literature reviews, medical professionals evaluating clinical evidence, data scientists building evidence-grounded ML models, and AI developers building research-augmented applications. Best suited for individuals and small teams (academic or industry) who need structured access to experimental data, not bulk enterprise users or organizations requiring offline/self-hosted solutions.