humanovo
Vol. I · No. 01 · Documentation

The table of contents
for everything we ship.

Read in the order written, or skip to the chapter you need.

Read first

  1. Manifesto

    Why humanovo exists. The thesis a researcher should agree with before investing time in the tool. Five hundred years of looking, rendered into a paragraph about the next chapter of that practice.

  2. Provenance — the citation roundtrip

    The methodology section. How every claim is verified against CrossRef and NCBI before it lands on your screen, what failure modes we explicitly catch, and how the per-hypothesis audit log makes any rendered claim later-replayable byte-for-byte.

  3. Atlas — the plate sources

    Every Renaissance plate used on the site, with full attribution to the holding institution, license confirmation, and a short editorial note on the choice. The lowest-marketing-cost trust signal we ship.

Getting started

  1. Install the desktop appIn progress

    macOS (Universal), Windows (x64), and Linux (deb / rpm / AppImage). Codesigned and notarised on macOS; signed with our EV cert on Windows; signed AppImage on Linux. The download links live on the homepage; the install instructions per-OS will live here when the binary is in public beta.

  2. Your first hypothesisIn progress

    A walkthrough — pose a question, point us at the corpus you care about, and read the resulting hypothesis trace. Includes how to interpret the audit log, what each pipeline stage scores, and where the gold-standard examples live.

  3. Loading your private corpusIn progress

    How to upload PDFs, sync from Zotero, point at a Notion / Obsidian / Roam workspace, or connect a Google Drive folder. Per-source ingestion notes, OCR caveats for scanned PDFs, and the privacy story for each.

The pipeline

  1. The 12 adversarial stagesIn progress

    What each of the twelve pipeline stages does, what it scores, and how to read the per-stage trace. Generation, evidence grounding, mechanism extraction, contradiction-search, counter-argument, revision, and the rest.

  2. The biomedical sources we readIn progress

    PubMed Central, bioRxiv, medRxiv, ClinVar, Reactome, ChEMBL, OpenTargets, GTEx, the Human Protein Atlas, ICTRP, and the rest of the 60+ open-access sources humanovo holds in one corpus. Each with the indexing cadence, license terms, and the canonical URL.

  3. How citations are scored

    Beyond the binary verified/unverified roundtrip — the cohort-concentration score, the predatory-source flag, the AI-generated-paper heuristic, the retraction-watch lookup, and how each surfaces in the UI as a chip you can click to read the rationale.

API & integrations

  1. API referenceIn progress

    REST + Webhook surface. Authentication via API keys with scoped permissions, rate limits per tier, idempotency keys for pipeline starts, and pagination conventions across list endpoints. Reference shipping with the public-beta API release.

  2. Webhooks & eventsIn progress

    Subscribe to hypothesis-completed, citation-verified, retraction-detected, audit-log-committed events. Signed payloads (HMAC-SHA-256), at-least-once delivery, replay endpoint for last 30 days. Spec lands with the API reference.

  3. ExportsIn progress

    Hypothesis exports in Markdown, BibTeX, RIS, and JSON. Audit log exports as JSON-Lines plus the Merkle commit hashes. Workspace bulk export as a tar.gz suitable for re-import or archival.

Operating the service

  1. Pricing & billing

    Four tiers, capped compute, no surprise bills. Invoices, currencies, taxes, and the academic discount mechanic.

  2. Privacy policy

    Our four privacy commitments and the operational detail behind each. Encryption, regional pinning, the inference-vendor roster, and the data-subject rights we honour globally regardless of jurisdiction.

  3. Terms of service

    The contract between humanovo and the people who use it. Plain-language acceptable-use, IP, liability, and dispute-resolution clauses.

  4. Status

    Live platform status, refreshed every 30 seconds: database, vector store, billing, and the 62-source biomedical orchestrator. Reads /api/v1/health/full directly. Institution-tier SLA terms remain in the contract; the status page itself is everyone's view of operational health.

  5. Support

    Email hello@humanovo.net. Response within 24h on Researcher tier, 4h on Lab, 2h on Institution.

A chapter you need that isn’t here yet? Email hello@humanovo.net. We will write it next.

Read the manifesto →