The tool offers deterministic verdicts and integrates with authoritative sources like PubMed and ClinVar, making it a reliable option for verifying biomedical evidence and detecting retractions.
Questions about biosingularity
What is biosingularity?
Biosingularity provides policy-driven biomedical evidence screening for manuscripts, research artifacts, and structured AI propositions. It checks references, gene-disease associations, drugs, clinical trials, and cell lines for issues like retracted sources and problematic cell lines. The service is intended for biomedical AI teams, researchers, journals, research-integrity groups, pharmaceutical companies, and biotechnology organizations.
How does the verification process work step by step?
The platform uses three explicit stages to separate entity matching from verification. First, it pulls out cited papers and gene/drug/disease claims. Second, it resolves each claim against primary sources like ClinVar and gnomAD. Third, it generates a verdict with an exact receipt for every finding.
What specific items can be checked using the platform?
Users can check specific items such as a DOI, gene, drug, NCT trial ID, or a free-text claim. The system auto-detects the type of input and resolves it against primary sources like OpenFDA and PharmGKB. It also includes a free text experimental mode that allows review or reject decisions only.
How can users get started or access the service?
The service offers a free tier for academic use and provides a pip install package for integration. It also includes an MCP connector available on request for broader application integration. There is no mention of a formal booking or ordering process beyond using the website or API.
What are the reported performance metrics for the ACMG frequency rule?
The platform claims 100% precision on the ACMG frequency rule behind its reject function. It successfully flagged 5,179 common variants while ensuring 0 truly-pathogenic variants were wrongly rejected. The F1 score is 0.84 with a recall of 0.97 in the in-silico consensus.
How does the system ensure deterministic and consistent results?
The system operates deterministically, meaning the same input always yields the same verdict without using any models in the verdict path. This ensures bounded integrity failures are caught before biomedical AI output ships. Every finding includes an exact receipt for transparency.
Which primary sources are used for resolving claims?
The platform integrates with a wide range of primary sources including ClinVar, gnomAD, dbNSFP, eQTL/GTEx, Open Targets, and DisGeNET. It also accesses GWAS Catalog, STRING, ChEMBL, and the PubMed retraction index. These sources are used to resolve claims against primary data.
What does the Evidence Explorer feature do?
The Evidence Explorer feature searches direct OpenAlex and the SemOpenAlex knowledge graph together. It provides source-level fallback and an explainable rank for every paper found. This discovery tool is not a verdict but helps find candidate evidence for review.
What specific checks can be selected for a single item?
The platform allows users to pick specific checks such as gene + disease, drug label with FDA boxed warnings, or trial status with ICMJE compliance. It also includes a reference disputed check that queries scite for disputes. Users can also check cell lines for misidentification issues.
Who is the intended audience for this service?
The service is designed for teams shipping biomedical AI, researchers checking references, and journals or research-integrity groups. It also serves pharmaceutical and biotechnology organizations needing to verify structured claims. No other specific target audiences are mentioned.
What features are available for testing and exporting results?
The platform includes a demo depth feature that allows users to edit and re-run checks with their own inputs. It supports full variant and interaction depth via CLI, API, or MCP. Users can also copy the verdict as markdown or copy the link to the specific check.
What are the specific modes for making automated decisions?
The system uses a typed v2 contract to make ALLOW, REVIEW, or BLOCK decisions automatable. It also includes a free text mode for experimental review and reject only scenarios. A clean entity lookup cannot prove an arbitrary proposition, so the typed contract is recommended for automation.
How does the platform compare with other tools or what benchmarks are available?
The platform includes a benchmark link for users to see how it compares with other tools. It also provides a discovery tool that is not a verdict but helps find candidate papers. The system resolves every claim against ClinVar, gnomAD, and other primary sources deterministically.
What is the current rating and how many reviews exist?
The Trustburn rating is 3.4 out of 5 from 5 reviews. The platform claims 100% precision on the ACMG frequency rule behind reject. It also offers a free tier for academic use and an MCP connector on request.
What is the official website of biosingularity?
The official website of biosingularity is biosingularity.ai.
Post your question publicly — other visitors, and whoever manages this page, can answer. It will be visible on this page.
Want to hear about this company again? Follow it and we’ll email you when a new review appears.
Based on reviews, listing data and the company's website. Verify anything important with the company directly.
The deterministic nature of the verdict path is a significant strength for teams requiring consistent results without model variance. The ability to screen manuscripts for bounded integrity failures, such as retractions and problematic cell lines, addresses a genuine pain point in biomedical research. The integration with primary sources like ClinVar and ChEMBL adds substantial credibility to the findings. However, the distinction between the free browser demo and the full-depth checks via CLI or API remains somewhat ambiguous. It would be helpful to have clearer documentation on which features are available in the free tier versus the paid API access. This clarity would make it easier to decide if the tool meets our specific workflow needs before committing to a larger integration. The focus on a false green as the worst failure is a prudent design choice that aligns well with rigorous scientific standards.
The free browser check runs live, which makes immediate verification of gene and drug claims very convenient for quick reference audits.
so i was looking at biosingularity.ai to see if it could help with our reference checking workflow. the site talks a lot about deterministic verdicts and catching retractions before pubmed, which sounds neat. but honestly, the whole page is just a giant wall of text that repeats itself three times. i cant tell if this is a free browser tool or a paid api. they mention a 493 gb datalake and cli access but dont say how much that costs or if there is a free tier. without any clear pricing or a breakdown of what is included in the basic access, it feels like i am just guessing. i would love to know if i can just try a few doi checks before committing to anything, but the site does not say. it seems like i might need to contact sales or request access just to find out the basic details. until then, it is just a lot of jargon about primary sources and typed predicates with no real insight into the user experience or cost.
The preflight checks look incredibly sharp for catching retracted sources and problematic cell lines. However, the website lists CLI, API, and MCP access without clearly stating the pricing tiers. From the information available, it is hard to determine if this tool fits a small research budget or requires enterprise-level spending.
Questions & Answers
No questions yet — be the first to ask.
Sign in to post
Your text is saved — you'll come right back to post.
Don't have an account? Create one
News & Press
No news yet. This is where the business shares its own updates and press.
Photos
No photos yet. Photos shared by reviewers will show up here.
About biosingularity
biosingularity provides policy-driven biomedical evidence screening for manuscripts, research artifacts, and structured AI propositions. It checks references, gene-disease associations, drugs, clinical trials, cell lines, and textual claims for issues such as retracted sources, problematic cell lines, trial-reporting gaps, and contradicted claims. Findings include exact receipts, while items outside a declared policy are held for review. The service is intended for biomedical AI teams, researchers, journals, research-integrity groups, pharmaceutical companies, and biotechnology organizations.
- Website
- biosingularity.ai
- Categories
- Computer software, Research, Biotechnology
