FAQ
Short, direct answers to the questions researchers ask most before trusting BioSkepsis with their literature work: what the agent does, how it grounds claims, what it covers, how your data is handled, and how plans and access work.
#What BioSkepsis is and isn’t
#What is BioSkepsis?
BioSkepsis is an agentic AI research assistant for the life sciences. You ask a biomedical or life-science question in plain language, and the agent refines the query, runs categorized searches across 40M+ papers, screens results for biological relevance, reads the strongest studies in full text, and returns a synthesis where every claim is grounded in a traceable citation. The goal is to compress the path from a question to a cited, evidence-backed answer.
#Do I have to configure or drive the search myself?
No. The agent is guided by default: it plans the research lens, builds the search queries, selects sources, and produces the first analysis without manual setup. Advanced users can step through and confirm each stage, but that is optional: the agent runs the workflow end to end on its own.
#Is BioSkepsis a replacement for reading the papers?
No. It is an accelerator, not a substitute for scientific judgment. BioSkepsis surfaces the relevant literature, extracts methods and mechanistic detail, and links every statement back to its source passage so you can verify it. Final interpretation, especially for clinical or publication decisions, remains yours. Every claim carries a citation precisely so you can check the underlying study.
#How is it different from a general-purpose chatbot?
General assistants reason mostly from training data and can fabricate citations or misstate what a paper says. BioSkepsis instead retrieves from a continuously updated life-science corpus through a biology-native knowledge graph, reads full-text papers rather than abstracts, and grounds each claim in a specific source. When evidence is insufficient, it declines rather than filling the gap with plausible text.
#Accuracy and hallucination
#Does BioSkepsis hallucinate?
No retrieval-grounded system can honestly promise zero hallucinations. What BioSkepsis does is make them auditable. The reasoning engine is constrained to the published life-science literature and your own uploaded data; every claim carries an in-text citation to the passage it came from; and the system explicitly declines when the evidence is thin instead of inventing an answer. The practical result is measurably fewer citation errors and an evidence trail you can inspect.
#Does it read full papers or just abstracts?
Full text wherever it is openly available. The agent reads the complete paper (methods, results and discussion) when it can retrieve it, so it can extract experimental controls, caveats and mechanistic detail that abstract-only tools miss. Not every paper is openly available in full text: where only the abstract can be read, the study is still used, and its citations are marked differently from full-text-verified ones so you can see how deeply each claim was checked.
#What happens when the evidence isn’t there?
The agent says so. Rather than generating a confident-sounding guess, it reports that the literature is insufficient or conflicting for that specific question. An explicit “not enough evidence” is treated as a correct answer, not a failure.
#Can I interpret my own experimental results?
Yes. You can describe bench observations or upload notes, and the agent maps them against known pathways and mechanisms in the literature, explaining where your findings align with or contradict published evidence. This is interpretation grounded in citations, not speculation.
#Sources and coverage
#What literature does BioSkepsis search?
The agent searches a curated corpus of 40M+ papers spanning Biology, Medicine, Pharmaceuticals, Biotechnology, Agricultural & Food Sciences, Veterinary Science, and Environmental Sciences. Coverage runs from 1931 to the present and is refreshed weekly so your research stays current with the published record.
#What publication types are covered?
A broad range: journal articles, reviews, clinical trials, meta-analyses, studies, case reports, conference papers, datasets, editorials, letters and comments, news, books, and book sections.
#How does it find relevant papers rather than just keyword matches?
Retrieval is biology-native. BioSkepsis builds a knowledge graph linking papers through Gene Ontology (GO), MeSH terms, specific genes, and domain keywords, so it retrieves studies by true biological relevance rather than surface text similarity or raw citation counts. This is what lets it connect related mechanisms even when titles and abstracts share no common wording.
#Can I keep up with new work in my area?
Yes. A personalized Research Feed uses embeddings of the papers you save to surface newly published work that fits your research context, and you can receive daily, weekly, or monthly digests so relevant new studies reach you without manual searching.
#Data and privacy
#Does BioSkepsis train on my queries or documents?
No. EFEVRE TECH LTD does not use your prompts, uploaded documents, AI outputs, or any other user-provided content to train, fine-tune, or evaluate our own machine-learning models, and never sells or publishes your raw prompts, queries, or uploaded documents. Producing an answer does send your prompt and any attached documents to the AI provider that serves your plan. Today that provider is Z.AI, operated by JINGSHENG HENGXING TECHNOLOGY PTE. LTD and processing that content in Singapore: its data processing addendum for API services states that it does not store the content you provide or generate, because the content is processed in real time and is not saved on its servers, and its published policy does not say whether it trains models on API content, so we make no promise on its behalf. Two things are separate from that: we may license anonymised, aggregated research-trend statistics that cannot reasonably be traced back to you, and inclusion of your professional profile in directories licensed to life-science organisations would need your separate opt-in. There is no opt-in control in the Service yet, so no profile is included. Your conversations, Sources, and uploaded notes are private by default; sharing is opt-in and controlled by you. Full details are in the Privacy Notice.
#Is BioSkepsis GDPR-compliant?
GDPR compliance is not a badge a supplier can award itself, so here is what is in place and what is not. The data controller, EFEVRE TECH LTD, is established in the EU and acts as controller under Article 4(7) GDPR for personal data processed through the platform, and you can access, rectify, restrict, port, object to, or delete your personal data at any time. Your account, conversations, and Sources are held on our own dedicated server inside the European Economic Area, and our sub-processors are named in full in the Privacy Notice, among them Z.AI in Singapore for the model that answers your question, Clerk for sign-in, Stripe for payments, Hetzner for hosting, Cloudflare for content delivery, Resend for email, and Sentry for error monitoring. Traffic to and from the platform is encrypted in transit with TLS 1.2 or newer, HSTS, and a Content Security Policy, and our vulnerability disclosure route is published at bioskepsis.ai/security-policy. Not in place today: encryption at rest, including for backups; audit logging of who read or exported what; and any external certification or commissioned penetration test.
#What are my rights over my data?
As a GDPR-protected user you can access, rectify, restrict, port, or object to the processing of your personal data. Account deletion in your account settings removes your conversations, Sources, attachments, profile, and feeds, and it cannot be undone. Some records outlive it today, including saved clusters, API keys, push subscriptions, product analytics events, the request log, an eLabNext connection, and the credit and billing records, and free-plan usage counters are kept deliberately. To have the remainder erased, or to exercise any other GDPR right, contact [email protected].
#Are my conversations shared with anyone?
Not unless you choose to share them. Every conversation, source list, and uploaded note is private by default. Sharing happens only through a private invite link or a public publish action that you initiate and control.
#Plans and access
#Is there a free plan?
Yes. The Basic plan is free, with no credit card required to sign up. It includes semantic search across 40M+ papers, the literature landscape graph, hypothesis and methodology generation, all smart-select filters, personalized Research Feeds with email alerts, and a monthly usage allowance for full, cited research runs. When you outgrow the free limits, you can upgrade at any time.
#What paid plans are available?
Four paid tiers. Plus (€18/month or €180/year) suits master’s and PhD students, with 3x the Basic monthly usage allowance, no monthly cap on the four analyses, and credits. Pro (€75/month) is built for postdocs, PIs, and healthcare professionals, with 13x the Basic monthly usage allowance and unlimited saved clusters. Team (€125/month for 3 seats, then €37 per extra seat, minimum 3 seats) gives each seat 9x the Basic usage allowance and adds shared billing and volume pricing. Organization is €150 per seat per month for pharma, biotech, and academic institutions, and every seat carries the Pro usage limits. Annual billing is published for Plus; for annual Pro, Team, or Organization pricing, write to [email protected].
#What does a “research brief” count as?
A research brief is one full agent run: literature search, relevance screening, and a cited synthesis. Your monthly usage allowance is a token budget, expressed as a multiple of the Basic allowance (3x on Plus, 13x on Pro), so how many full runs it covers depends on question complexity, follow-ups, and response length. On Team plans, allowances are granted per seat.
#What are credits?
Credits let Plus, Pro, and Team users keep working past their monthly allowance instead of hitting a hard stop. You buy a balance in Billing settings, and research beyond your included allowance draws it down; the current credit price is shown there before you buy. You can turn on automatic top-ups and set a monthly top-up limit, and expiry, refunds, and billing are governed by our Terms. On Team plans the owner buys credits and the whole team shares one balance. Basic does not include credits: on Basic you wait for the next month or upgrade.
#What does the Organization tier add?
Organization is €150 per seat per month, and every seat carries the Pro usage limits. What it adds over Team is institutional administration rather than a different engine: a single contract, a single invoice, one member list for everyone in the organisation, onboarding and training, and priority support in Cyprus business hours, with a first response within 4 hours for the most severe problems. Single sign-on and admin controls for the Organization tier are in build rather than available today; write to [email protected] to ask about the timeline.
#Can I cancel anytime?
Yes. You can cancel any paid plan in one click from your Billing page. Your plan stays active until the end of the current billing period, after which you revert to Basic with no further charges. Saved conversations and Sources remain accessible under the free plan’s limits. Annual plans are not pro-rated mid-period but can be cancelled at any time to stop the next renewal.