☸️SAIMSARA
AI Systematic Review & Meta-Analysis Platform

×

How to Use SAIMSARA

1. If you do not subscribe to SAIMSARA Premium, you can start directly from point 9.
2. For the advanced workflow, click the Login button.

   

3. Follow the Subscribe link.

   

4. Choose your plan — a 1-week pass (one-time) or a monthly subscription.

   

5. Input your email and proceed with the payment.

   

6. Log in with your email address.

   

7. Receive a notification email with access credentials.

   

8. Enter the PIN or click the direct login link from your email.

   

9. 🎉 Premium access enabled! Enter your keywords and click Search.

   

10. Wait for the first batch of AI analysis — 25 papers in Free or 50 in Premium.

    

11. After the search is done, track the total number of articles and the % analyzed. To continue the search, click Next Articles or Autorun for online autosearch, or choose Offline Autorun to run the search offline and receive the paper by email. Note: Autorun is available to Premium subscribers only

    

12. Click Craft paper to craft a review paper any time. You’ll be notified if the search is incomplete. For full evidence, extract all available articles.

    

13. The crafted paper appears in the chat. A session file maps each numbered citation to references.

    

    

14. Premium users get extended visuals and an individual web link to the paper (web page).

    

    

×

☸️SAIMSARA

Systematic, AI-powered, Multilayer, Scientific Article Review Agent

What it does

☸️SAIMSARA is an autonomous multilayer AI research agent for rapid, PRISMA-compliant screening of biomedical literature. It analyzes records the trusted database Semantic Scholar (AI2) — covering over 230 million research papers — , filters out non-original and irrelevant studies, and synthesizes the remaining evidence into:

• Clear numerical results
• Main research topics (practical standpoint)
• Study limitations
• Future research directions

All steps and extracted data are stored in a unique Session URL, ensuring full transparency and reproducibility.

Outcome Sentiment Meta-Analysis (⛛OSMA)

☸️ SAIMSARA includes an LLM-driven evidence synthesizer that reads each study’s Main Results and Statistics and groups findings into buckets with ΣN (sample-size) weighting:

• Head-to-Head: “A vs B” comparisons classified as Favours A, Favours B, or Neutral (no clear difference).
• Effect-of: Predictor → Outcome assessed for Beneficial for patients, personnel, populations, Harmful, or No clear effect.

The ⛛OSMA Triangle visualizes these buckets, with every segment traceable back to the original studies. ⛛OSMA provides a rapid, human-centric layer of evidence mapping that complements traditional meta-analysis.

Why it matters

To evaluate a project, validate an idea, or identify promising research or investment directions, you need deep analysis of existing data. Without it, narratives are often incomplete, leading to under- or overestimation of a project’s value. SAIMSARA delivers fast, transparent, evidence-based insights that support stronger decisions in research, industry, and academia.

More than a search engine or chatbot

• Not Google, Yahoo, DuckDuckGo, or a generic LLM interface.
• Does not rely on its own memory or guesses—works from ~50 million peer-reviewed biomedical papers.
• Every output is a traceable, referenced piece of evidence.

Examples

☸️SAIMSARA Archive

Is SAIMSARA free?

Yes. You can use the Free version without charges and without registration. It includes the following features:

• AI agent: Gemini 2.0 Flash Lite
• Database: PubMed (~39 million records, isolated search)
• Batch size: 25 records per batch, with manual switching between batches
• Tokens: ~8,000 output tokens per batch for extraction and summary
• Session link: Each review session generates a unique link with review text, references, and statistical figures

Maximal Evidence with Premium

• One-click Search→Paper workflow
• Full offline autorun of SAIMSARA sessions, with automatic delivery of the generated review paper to your email.
• Top reasoning model -Gemini 2.5 Pro- for summary
• Database ~50 million articles from PubMed + Europe PMC
• Synthesis Speed - 100 papers per minute
• 50 abstracts per batch
• 65k output tokens
• Enhanced graphical summaries
• Review paper as a shareable web page (URL)
• DOI assignment on request
• Priority support
• and more...

Note: App batch sizes and token budgets shown above reflect current defaults and can be tuned in code constants if needed.

Follow us

×

Version history

• Version 3.6 (09.10.2025): Switched to Semantic Scholar (AI2) as the primary search engine — JSON-based semantic indexing replacing XML pipelines. Marks the transition toward fully ML-native literature retrieval.
• Version 3.5 (03.10.2025): Introduced Outcome Sentiment Meta-Analysis (⛛OSMA) — automated evidence bucketing (benefitial, harmful, or neutral for patients or study participants) with sample size ΣN weighting, plus ⛛OSMA Triangle visualizations.
• Version 3.0 (13.09.2025): Added automated search in the Premium version — a fully agentic AI-driven workflow from initial search to paper creation. UI improvements: DOI link visualization with embedded metadata preview and an inverted (high-contrast) dark theme. Added visual summary with Figures.
• Version 2.5 (10.09.2025): Updated hard keyword filter prior to sending to LLM. Updated Reference logic.
• Version 2.0 (03.09.2025): Added paid version with PMC + PubMed Search and Gemini 2.5.
• Version 1.5 (02.06.2025): Gemini update to 2.0 Flash.
• Initial (20.02.2025): PubMed only, Gemini 1.5 Flash.

×

☸️ Archive

×

Contact us

Your Email Address:

Your Message:

Leave this field empty:

Submit