CONTEXTVM · A WORLDVM EXPLORATION
Read the corpus.
Discover its ontology.
ContextVM reads any corpus of data and documents the ontology emerging from it: the concepts, fields, and relationships that give the data its shape.
Bring research, notes, records, manuals, or agreements. Explore their shared vocabulary, follow the evidence, and export an ontology you can review and build on.
Try a research corpus or explore service agreements as one example. Your files stay in your browser.
field-notes.md · line 3
## Sampling
Method: Weekly water samples
## Equipment
Instrument: Temperature logger
Calibration: Monthly
Emerging concepts
Sampling2 sourcesheading
Equipment2 sourcesheading
studies2 recordsJSON
people2 recordsJSON
Observed relationships
studies → matches investigator → people
HOW IT WORKS
From a corpus to an ontology, with evidence.
-
1
Read the corpus
Import PDFs, Markdown, text, JSON, and CSV together. Keep readable Markdown and source locations alongside the data.
-
2
Find the vocabulary
Concepts and fields emerge from the corpus’s headings, collections, and values. Start with the data’s own structure and vocabulary.
-
3
Connect the concepts
Explore shared fields, nested structures, and references between records. Follow each candidate relationship back to its source.
-
4
Document what emerges
Export the concepts, observed types, relationships, and evidence as Markdown. Review the meaning and use it to shape a shared model.
WHAT IT’S FOR
For any body of knowledge.
Research collections, customer records, operating manuals, or agreements: each corpus carries a vocabulary of its own. ContextVM makes that structure visible.
- Discover the data modelSee the concepts and fields emerging from documents and structured records without starting from a domain schema.
- Bring scattered knowledge togetherFind repeated concepts across sources and inspect the evidence behind their relationships.
- Document the ontologyExport a vocabulary with observed types and source citations that a team can review and refine.
- Example: service agreementsExplore a specialized service agreement analyzer with cited terms, cross-contract search, benchmarks, and queries. One use case for ContextVM.
Honest status
- The product direction is any corpus. This browser exploration imports text-based PDFs, Markdown, text, JSON, and CSV.
- General discovery uses structural rules, with no predefined domain schema. Candidates come from headings, collection names, fields, nesting, and identifier matches; semantic inference needs a model and review.
- Scanned PDFs need OCR. Data stays in this browser tab; export it before closing.
- The separate service agreement example uses a fixed 23-term schema and 1,240 synthetic agreements to demonstrate a deeper domain workflow.