MAMaxArc Medical Library™Health information and medical learning
Care assistanceSelected services are in controlled preview; inactive functions are clearly marked where they appear.
Skip to main content

MaxArc Biomedical Research Intelligence™

BioResearch

A biomedical research intelligence platform with AI-assisted search, evidence discovery, and locally indexed research.

Research records are acquired as versioned snapshots, validated, normalized and de-duplicated into the MaxArc canonical research index, and searched locally. A citation is not an article body, and an abstract being available upstream is not permission to reuse it — every record states both separately.

Educational information only. This explains health information to discuss with a qualified clinician; it is not a diagnosis, prescription, or treatment. Medical Information & Safety.

Local index
Available
Upstream sync
Optional / off
Last validated snapshot
maxarc-fixture-2026-01
Source
MaxArc synthetic fixture set
Research records indexed
12

No real upstream call is made. BioResearch answers from the MaxArc local research index, which is how it is designed to run. Optional upstream synchronization is governed by the PUBMED_LIVE real-effect flag, which controls real provider traffic only: with it off, BioResearch stays visible, local search keeps answering, and the Owner can still test every level it supports. Open the Owner readiness surface for this module.

These 12 research records are not encyclopedia concepts. The Medical Encyclopedia holds 28,460 canonical concepts, counted separately and unchanged by anything on this page. All 12 records in this foundation are MaxArc synthetic fixtures — no bulk PubMed or MEDLINE ingestion has been performed, and no article body is stored.

Sources

Every source enters through a provider-neutral adapter that records its own identity, release, rights basis and redistribution restriction. Provider payload shapes stop at the adapter.

MaxArc synthetic fixture set

Owner test
Adapter:
maxarc-fixture-adapter 1.0.0
Acquisition:
fixture · incremental updates not supported by this source
Rights basis:
metadata-public
Redistribution:
No provider content is included, so nothing here grants rights over real source material.
Attribution:
Synthetic records authored by MaxArc. Not derived from any provider.

Search backend

Owner test

Search is asked of a provider-neutral backend contract, not of an in-process record list this page holds: maxarc-in-memory-research-index 1.0.0 answers 12 records from maxarc-fixture-2026-01 with no upstream provider. This fixture-scale implementation scans the whole index per query, which is why a PubMed-scale index needs a different implementation of the same contract rather than a different product surface.

PubMed / MEDLINE

Integration pending

A PubMed/MEDLINE adapter is the next layer, not this one. Its citation metadata will enter through the same pipeline, and citation metadata never implies reusable article full text. Optional real provider traffic is gated by PUBMED_LIVE.

Evidence types

Publication types held in the local index. They describe what a record is; they are not an evidence grade, and formal guideline mapping and evidence grading are a separate layer.

Evidence grading and guideline mapping Not yet implemented

Source freshness and snapshot status

A snapshot is promoted only when every gate passes. A malformed or failing snapshot is quarantined and the previously validated index keeps serving, so BioResearch degrades to the last known-good index rather than to nothing.

Snapshot
maxarc-fixture-2026-01
Source version
fixture-1.0.0
Acquired
2026-01-15T00:00:00Z
Stage
promoted
Raw records
12
Normalized
12
Rejected
0
Quarantined snapshots
0

Raw checksum sha256:9f2c0a1d4b7e5c3a8d6f1b0e2c4a7d9f3b5e8c1a0d2f4b6e8c0a2d4f6b8e0c2a

  • source-identity — passed · index, adapter and snapshot all declare maxarc-fixture
  • acquisition-complete — passed · fixture of maxarc-fixture-2026-01 recorded
  • checksum-recorded — passed · sha256:9f2c0a1d4b7e5c3a8d6f1b0e2c4a7d9f3b5e8c1a0d2f4b6e8c0a2d4f6b8e0c2a
  • schema-validated — passed · schema maxarc-fixture-rows/1 accepted
  • normalization-complete — passed · 12 normalized, 0 rejected
  • rights-metadata-present — passed · every record carries an attribution and an indexable rights state
  • dedup-complete — passed · 12 distinct records, 0 duplicates collapsed
  • record-count-plausible — passed · 12 records, declared 12
  • previous-index-preserved — passed · first promotion for this source

Research workspace and saved research

Named honestly: each capability below carries its readiness state rather than a working-looking interface.

Research workspace

Not yet implemented

Per-project collections, appraisal notes and export are designed and not built. Nothing here is presented as working.

Saved research

Not yet implemented

Saving a record needs an identity-bound store and retention policy. The foundation deliberately keeps no user state.

AI-assisted query expansion

Preview

Deterministic field-qualified search is live against the local index. Assisted query expansion and evidence summarization remain designed only, and would never produce a clinical recommendation.

Knowledge Spine linking

Integration pending

Typed relationships between a research record and a canonical Medical Knowledge Spine concept exist as interfaces. No records are linked in this foundation, and a link would need human review before it informed any presentation.

What BioResearch does not do

  • It does not diagnose, prescribe, or make a patient-specific clinical decision, and it holds no patient data.
  • It does not store article full text, and it does not treat citation metadata or abstract availability as permission to redistribute a body.
  • It does not count research records as Medical Encyclopedia concepts, and it does not change the encyclopedia’s 28,460 canonical concepts.
  • It does not require a live upstream provider to answer a research search.