What the tool is
SotaSearch.ai is a privacy-first, evidence-driven research assistant that runs entirely in the browser (WebGPU/WebLLM). It can also use local Ollama or user-provided keys for OpenAI/OpenRouter. Core features include:
• Paper Map: detects literature gaps/future directions and drafts novel paper abstracts grounded in the reviewed corpus.
• Semantic & citation graphs: interactive mapping of topics, clusters, and citation relationships.
• Academic search integrations: PubMed, Semantic Scholar, Cochrane, Scopus, Google Scholar; citation export (APA/MLA/IEEE/ISO/BibTeX).
• Local-first privacy posture: 100% local processing, zero data collection, local storage; offline-capable.
Intended users / use cases
• Researchers and students performing systematic or narrative literature reviews.
• Labs and groups running paper clubs who want structured summaries + gap identification.
• Professionals needing evidence synthesis with strong privacy guarantees.
Try it / status
• Site & demo: https://www.sotasearch.ai/
• Status: beta; Chrome Web Store listing forthcoming.
• Setup docs cover WebGPU in-browser models, optional Ollama, and optional BYO API keys.
Policy & community commitments
• I will abide by Reddit’s Rules and this subreddit’s specific rules.
• No cross-posting spam, no vote manipulation, and no private outreach.
• If a monthly self-promotion/“projects” megathread exists, I will post there instead.
• Feedback—especially critical or comparative— is explicitly welcome. If you prefer that posts about tools include benchmarks, I’m happy to provide transparent comparisons and limitations.
If there is interest, I can share a reproducible mini-demo (public abstracts only) showing Paper Map → gap bullets → draft abstract, plus a small semantic/citation graph for a current topic.
Moderator note / disclosures (please read first)
— Affiliation: I help promoting SotaSearch.ai (currently in public beta) for free.
— No tracking, no referral links, no DMs or solicitations. I will not ask for upvotes or messages; discussion stays in-thread.
Intended users / use cases • Researchers and students performing systematic or narrative literature reviews. • Labs and groups running paper clubs who want structured summaries + gap identification. • Professionals needing evidence synthesis with strong privacy guarantees.
Try it / status • Site & demo: https://www.sotasearch.ai/ • Status: beta; Chrome Web Store listing forthcoming. • Setup docs cover WebGPU in-browser models, optional Ollama, and optional BYO API keys.
Policy & community commitments • I will abide by Reddit’s Rules and this subreddit’s specific rules. • No cross-posting spam, no vote manipulation, and no private outreach. • If a monthly self-promotion/“projects” megathread exists, I will post there instead. • Feedback—especially critical or comparative— is explicitly welcome. If you prefer that posts about tools include benchmarks, I’m happy to provide transparent comparisons and limitations.
If there is interest, I can share a reproducible mini-demo (public abstracts only) showing Paper Map → gap bullets → draft abstract, plus a small semantic/citation graph for a current topic.
Moderator note / disclosures (please read first) — Affiliation: I help promoting SotaSearch.ai (currently in public beta) for free. — No tracking, no referral links, no DMs or solicitations. I will not ask for upvotes or messages; discussion stays in-thread.