PICO-SEARCH
A real-time literature search engine for clinicians and the general public — surfaces existing systematic reviews, meta-analyses, RCTs and guidelines in seconds, graded against the CEBM hierarchy. Your question is routed to the right clinical specialty, the literature is retrieved live, every study is graded against the CEBM hierarchy of evidence, and the findings are laid out in a PRISMA-style format with every claim linked to its source paper. We present the data. We don't invent, advise, or make claims.
Scope honesty. PICO-SEARCH is tier 3 of the evidence-review spectrum: real-time retrieval (seconds) → rapid review (days–weeks) → full Cochrane systematic review (6–18 months, dual-reviewer, PROSPERO-registered). We are the first tier. To escalate to a systematic review, export the search strategy + RIS file into Covidence, Rayyan or RevMan.
Nerd Burger + 24 Sliders
- Neurology & Nervous System
- Ophthalmology & Vision
- Ear, Nose & Throat
- Cardiovascular
- Respiratory
- Gastrointestinal & Hepatology
- Endocrine, Nutrition & Metabolic
- Renal & Urology
- Musculoskeletal & Rheumatology
- Dermatology
- Mental Health & Psychiatry
- Obstetrics & Maternity
- Gynaecology & Women's Health
- Men's Health
- Paediatrics & Child Health
- Geriatrics & Older Persons
- Haematology
- Oncology
- Infectious Disease
- Immunology & Allergy
- Emergency & Critical Care
- Rehabilitation, Pain & Palliative
- Public Health & Preventive Medicine
- Dental & Oral Health
How it works
Every search passes through four layers. First, the Nerd Burger — the coordinator AI — reads the question and identifies which clinical specialty it belongs to. Second, the relevant Sliders— twenty-four specialist agents, each trained in one clinical area (cardiovascular, respiratory, mental health, women's health, paediatrics, and so on) — run a live literature search across every major medical database: PubMed, Europe PMC, ClinicalTrials.gov, and the rest.
Third, every retrieved study is graded against the hierarchy of evidence — systematic reviews and meta-analyses at the top, expert opinion at the bottom. The strongest evidence is surfaced first, regardless of which study is most recent or most talked about.
Fourth, the findings are laid out objectively with every single claim linked to the original source paper. We do not interpret. We do not advise. We present the data. This is how a research librarian starts an evidence search — the architecture does the same, in under a minute. PICO-SEARCH retrieves existing systematic reviews, meta-analyses, RCTs and guidelines — it does NOT conduct a new systematic review (that is a 6–18 month dual-reviewer workflow). When an answer warrants the full SR treatment, export the search strategy + RIS into Covidence, Rayyan or RevMan and escalate from there.
Three ways to use it
Our aim: make clinical evidence searchable — so the data, not the noise, shapes health conversations.
Via MCP
PICO-SEARCH runs a Model Context Protocol server. Add it as a tool in Claude, ChatGPT, Gemini, Cursor — any MCP-compatible AI. When your assistant is asked a health question, PICO-SEARCH runs the actual literature search and returns graded evidence — so the AI quotes the data, not its training set.
MCP endpoint →Via API
PICO-SEARCH exposes a bearer-authenticated HTTPS API. Your patient-management software, EMR, or internal clinical tool can call it directly — the structured, cited, evidence-graded output streams back into the interface you already use.
Get an API key →Right here
Open picosearch.ai or nerdburger.ai. Type a question. Read the structured answer with every claim linked to the source paper. Two front doors, same brain: picosearch.ai renders the output as a PRISMA-style rapid evidence summary; nerdburger.ai renders it in plain English. Neither is a full systematic review.
↑ Try the search above