☸️SAIMSARA Chat
Scientific AI Agent
Mental Health grounded chat

Searches scientific papers and extracts evidence into standardized PRISMA-style scoping reviews, using major scientific libraries including:

• Semantic Scholar — 230M+ papers; retrieves the top 1k matches; best for broad discovery and rare targeted questions.
• Europe PMC — 47M+ records; supports full-scale search, including full text where available; best for maximum search depth.
• PubMed — 39M+ records; retrieves the top 10k matches; best for precise biomedical and clinical retrieval.

In experimental settings, SAIMSARA has generated reviews with ~20,000 traceable references in ~120 minutes — compressing months of manual literature work into a single session.

ISSN 3054-3991

Hub of machine-generated scoping review papers.

Papers assigned to domain-specific issues are checked for clinical integrity, reference quality, and scientific coherence by a World ID verified human editor.

Each issue is paired with its own dedicated AI agent, allowing readers to chat directly with the Journal.

SAIMSARA uses World ID to verify that editorial oversight is submitted by a unique human account. This adds a proof-of-human editorial layer above the multi-LLM paper-generation workflow and helps distinguish human-reviewed publications from machine-only outputs.

LLM interface enhanced with direct access to the SAIMSARA scoping review database.

In the Pro version, chat includes access to Deep Panel, a credit-based advanced mode in which Claude 4.6 Sonnet, ChatGPT 5.5, Gemini 3.1 Pro, and Grok 4.3 reason independently before a final judged synthesis is produced by Claude Opus 4.7.

SAIMSARA Chat can also use openFDA APIs for regulatory drug-label information and medical-device records, including dosage, contraindications, warnings, and indications.

SAIMSARA Chat is intended for scientific and educational use and does not replace professional medical advice.

API access to the SAIMSARA scoping review database, designed to enhance external generative AI systems with independent, large-scale, machine-generated, and human-reviewed medical and life-science information. The Evidence API can be integrated into research, education, clinical-support, or AI/RAG workflows.

API v1 is read-only and does not expose user data. We provide the evidence layer. You build the application. You remain responsible for its use.