Evidex: A Privacy-First Clinical Search Engine

A solo developer has launched Evidex, a new clinical search engine designed to address the high costs and pharmaceutical advertising prevalent in existing medical reference tools. The creator built the platform to assist resident physicians and their colleagues who struggle with the expense and speed of current market leaders like UpToDate and OpenEvidence.

The platform distinguishes itself through a unique technical architecture that prioritizes data freshness and privacy. By utilizing a real-time Retrieval-Augmented Generation (RAG) pattern rather than pre-indexed vector databases, Evidex ensures that users have access to the most current medical literature, including trials published on the same day. The system integrates advanced AI capabilities using Gemini 2.5 Flash and offers workflow tools such as SOAP note drafting and complex patient history analysis.

The development of Evidex stems from a specific gap in the medical technology market. The creator identified that current clinical search standards are often expensive, slow, or increasingly heavy with pharma ads. These barriers can hinder the workflow of medical professionals who require immediate, unbiased access to clinical data.

To counter these issues, Evidex was built as a clean, privacy-first alternative. The platform focuses on delivering a streamlined user experience without the clutter of advertisements. The core mission is to provide a reliable tool that respects user privacy while delivering high-quality medical information.

Key features of the solution include:

• Ad-free search environment
• Privacy-centric data handling
• Real-time access to medical literature
• Cost-free clinical search capabilities

Evidex employs a sophisticated Search-Based RAG pattern that diverges from traditional methods. Instead of relying on a pre-indexed vector database like Pinecone—which can serve stale data—the system implements a Real-time RAG pattern. This approach is critical in medicine where 'freshness' is paramount; if a new trial drops today, a pre-indexed store might miss it, but Evidex ensures the answer includes papers published today.

The technical workflow involves several distinct stages:

1. Orchestrator: A Node.js backend performs 'Smart Routing' using regex and keyword analysis to determine which external APIs to query.
2. Retrieval: The system executes parallel fetches to APIs such as PubMed, Europe PMC, OpenAlex, or ClinicalTrials.gov at runtime to retrieve the top ~15 abstracts.
3. Local Data: Clinical guidelines are stored locally in SQLite and retrieved via full-text search (FTS) to ensure exact matches on medical terminology.
4. Inference: Gemini 2.5 Flash processes the concatenated abstracts. Its massive context window allows for feeding distinct search results and enforcing strict citation mapping without latency bottlenecks.

Beyond simple search capabilities, Evidex includes a reasoning layer designed to integrate deeply into clinical workflows. This layer handles complex patient histories through a dedicated Case Mode and assists with administrative tasks by drafting SOAP Notes. These tools aim to reduce the documentation burden on physicians.

Regarding sustainability, the developer has adopted a unique business model. The clinical search functionality remains free for users. Monetization is planned through the sale of billing automation tools to hospital administrators at a later stage. This strategy allows the core medical search tool to remain accessible while generating revenue through adjacent administrative services.

The developer is currently soliciting feedback on two specific areas:

• Retrieval latency (fetching live APIs is slower than vector lookups)
• Accuracy of the synthesized answers