PUBLIC DATA
SmellBook's public data, free to download.
No licence has yet been granted for SmellBook-created datasets. They are publicly readable; reuse beyond citation and quotation needs permission until a licence is published. Licence notice.
The files below hold a deliberately limited selection of the website's public facts, in forms that programs and spreadsheets can read. The descriptions use some technical terms because they are written for developers and researchers. These files can only be read, not changed. No user-account, write or recommendation API is currently provided, and no service-level agreement is implied. They hold nothing private: no hidden graph data, no search-ranking weights and no recommendation logic.
Manifest
Schema version, endpoint list, exported fields and explicit private-field exclusions.
Catalog
The 200 published smell entries with IDs, names, other names, summaries, descriptors, status and revision metadata.
Taxonomy
Public categories, descriptors and public types with their canonical smell memberships.
Evidence
The public source registry and VERIFIED claim records.
Changelog
Structured public product and atlas change records.
Provenance
Explicit profile-to-claim-to-source evidence relationships with stable public identifiers.
Relations
Explicit unweighted editorial related-smell links.
Identifiers
Canonical public smell, claim and source identifiers with canonical URLs.
Identifier resolver
Read-only exact-ID resolution example.
Quality report
Structural integrity and maintenance diagnostics for the public atlas.
Quality issues
One hard structural issue per line; empty for a passing release.
Current snapshot
Current documented public snapshot metadata.
Snapshot registry
Formal public snapshot history from 2026-09-22 forward.
Latest snapshot pointer
Stable pointer to the newest registered public snapshot.
Catalog table
The public canonical catalog in spreadsheet-friendly CSV.
Catalog table
The public canonical catalog in tab-separated form.
Claims table
Verified public claim records in spreadsheet-friendly CSV.
Claims table
Verified public claim records in tab-separated form.
Sources table
Public source records in tab-separated form.
Citation library
Public source records in CSL JSON for citation managers.
BibTeX library
Public source records in BibTeX form.
RIS library
Public source records in RIS form.
Markdown bibliography
Human-readable public source bibliography.
Catalog stream
One canonical profile JSON object per line.
Claims stream
One VERIFIED claim JSON object per line.
Sources stream
One registered source JSON object per line.
Changelog stream
One public changelog record per line.
Provenance stream
Profile, claim and source relationship records as line-delimited JSON.
Relations stream
One unweighted editorial relation record per line.
Smell entity record
Example stable per-smell JSON record.
Claim entity record
Example stable per-claim JSON record.
Source entity record
Example stable per-source JSON record.
JSON-LD context
Schema.org mappings plus public SmellBook vocabulary terms.
Linked-data graph
Public smell, claim and source entities as JSON-LD.
Dataset bundle
Portable manifest for snapshot, resources, schemas, checksums and citation guidance.
Data dictionary
Field types, nullability and public semantic definitions.
Taxonomy facets
Category, descriptor and public-type membership records.
Facet memberships
One facet-to-smell membership per line.
Category record
Example category facet record.
Descriptor record
Example descriptor facet record.
Type record
Example public-type facet record.
Data lineage
Version-controlled inputs and their derived public outputs.
Reproducibility contract
Build commands, snapshot/checksum pointers and runtime non-dependencies.
Consumer contract
Read methods, CORS, caching, schema headers and compatibility semantics.
Error catalog
Stable public error codes and HTTP status meanings.
Consumer examples
Representative read-only requests and expected results.
Resource registry
Methods, media types, schemas, cache classes and stability metadata for public resources.
Search corpus
Searchable public fields and records without internal scoring weights.
Search corpus stream
One searchable public record per line.
Search field contract
Search field meanings and normalization semantics.
Cache policy
Public response cache classes and freshness semantics.
Capabilities
Current public support state and routes.
Support matrix
Capability status meanings and current boundaries.
Source lifecycle
Current lifecycle states for registered sources and VERIFIED claims.
Corrections
Correction notices SmellBook has published, each keyed to the permanent identifier of the record it corrects.
Rights registry
Current public/project/third-party reuse and ownership boundaries.
Attribution registry
Registered source attribution and publisher-rights guidance.
Evidence coverage
Profile/source/claim VERIFIED-linkage coverage without confidence scoring.
Vocabulary registry
Controlled public labels and definitions.
Privacy posture
Current runtime-aligned public privacy declarations.
Accessibility status
Source-level commitments and structural requirements.
Locale readiness
Current English-first language/localization posture.
Documentation map
Grouped public developer/policy documentation routes.
OpenAPI
OpenAPI 3.1 description of representative stable public read routes.
Snapshot deltas
Registered snapshot metadata/resource/count deltas.
Record fingerprints
SHA-256 fingerprints over selected public record fields.
Provenance chains
Explicit smell-to-VERIFIED-claim-to-source paths.
Provenance chain stream
One explicit evidence chain per line.
Deprecations
Current public deprecation registry and policy.
HTTP semantics
Methods, CORS, shared headers, media types and status meanings.
Retention policy
Public atlas traceability, snapshot and removal boundaries.
Ordering contract
Deterministic collection order and serialization semantics.
Canonical registry
Entity canonical URLs, route families and redirect policy.
Query contract
Documented query parameters, limits and error behavior.
Data minimization
Explicit public/private exclusions and exposure rules.
Security posture
Configured response protections and explicit non-claims.
Stability registry
Current route stability classes and field-change semantics.
Schema examples
Representative current public payloads selected from live data structures.
Evidence recency
Descriptive source publication-age metadata.
Editorial provenance
Public field-origin and third-party source metadata origins.
Download contract
Suggested filenames, media types and download boundaries.
Smell map topology
Deterministic visual coordinates plus explicit unweighted public relation edges.
Personal SmellBook local contract
Browser-local storage schema, limits, portability and privacy boundaries—not user data.
Applied scent contract
Public scent intents, sensory directions, space contexts, presence levels and safety/appropriateness boundaries.
Labs experimental export
Transparent EXPERIMENTAL representations and PREDICTED relation candidates with declared methods and boundaries.
Labs evaluation export
Deterministic run manifests, candidate explanations and descriptive coverage audits without confidence or performance scoring.
Labs review export
Read-only review packets, non-numeric rubric prompts and deterministic challenge sets with zero promotion effect.
Labs review context export
Claim-specific VERIFIED context, registered-source context, evidence-gap markers and provenance trails with zero relation-verification effect.
Labs review lenses export
Deterministic review-packet groupings and a descriptive category coverage matrix without ranking, confidence or quality scoring.
Labs change audit export
Deterministic SHA-256 fingerprints over declared public Labs fields with NOT_COMPARABLE historical semantics when prior fingerprints were not retained.
Labs audit snapshots export
Immutable retained audit-input sets beginning on 2026-09-29, with literal public input projections and SHA-256 digests.
Labs audit comparison export
Truthful comparison between retained snapshots, with descriptive ADDED, REMOVED, CHANGED and UNCHANGED states.
Knowledge graph export
Every entity with its permanent identifier and every typed relationship with provenance and evidence class, released through the rights gate with licence status and a sha256 digest.
Knowledge graph export stream
The same knowledge graph as header, node and edge records, one per line.
Permanent identifiers
The append-only registry of smellbook:{type}:{key} identifiers with sealed issuance blocks and retirements.
Entity model
Entity types with definitions and every entity record.
Relation types
Typed relationship vocabulary with semantics, direction, status and edge counts.
Evidence model
The six evidence classes, provenance mapping and the design-only contribution review model.
Data rights
Rights classes, field-level rights, source attributions and the dataset licence status.
Locales
English base language, the Spanish pilot names, review states, fallback and hreflang rules.
Source intake
Source intake workflow, queue, registered source rights and evidence gaps.
Entry standard
Publication criteria, development signals and the result for every entry.
Atlas quality baseline
The twelve entry checks, every entry's facts and findings with their resolutions, and the descriptor review. No score or ranking.
Entry pipeline
Editorial draft, review and publication workflow for new entries.
Entry identity
Name uniqueness, reviewed near-duplicates and the merge policy.
Evidence research agenda
Open research questions for entries without a VERIFIED claim.
Historical sources
Historical and archival sources: holdings, catalogue records, digitised copies, dates with their precision, and rights. Catalogue metadata identifies a source; it never verifies a claim.
Historical source worksheet
One row per proposed historical passage: catalogue record, digitised copy, location, the time it describes and rights. Registered passages are filled examples.
For institutions
How museums, archives, universities and teachers can use SmellBook: uses, identifier guidance, research bundles and offline tools.
Heritage record templates
Templates for object, archival-document, place and practice records, with worked examples from registered evidence and invented illustrative records.
Institution self-test kit
Example records and a LIDO export, some deliberately broken, with the result each must give.
Entry citations (RIS)
Every entry as an RIS reference with its record number, permanent identifier and revision.
Evidence bundle
VERIFIED claims, how they were checked, historical sources and molecules, with the SHA-256 of every part.
Olfactory heritage
Localities and traditions in the olfactory heritage atlas, with their identity references and check status.
Molecules
Odorant molecules named in VERIFIED claims, the role each claim gives them, and their verified PubChem, InChIKey, Wikidata and ChEBI identifiers.
Claim support
How each VERIFIED claim was checked against its source's own words; 2 could not be, because no abstract was available.
Citation verification
How every source citation was checked against the publisher's or catalogue's metadata, with the check runs.
Graph records
One knowledge-graph entity with its typed relationships: /data/graph/{type}/{key}.json.
Dataset releases
Every release of the knowledge-graph dataset: what changed, its digest, and how to cite it.
SKOS vocabulary
Smell descriptors (English and Spanish labels) and places by kind as SKOS concept schemes for thesaurus and collection systems.
Heritage record schema (draft)
Draft JSON Schema for olfactory heritage records that link to SmellBook identifiers. No institution uses it yet.
Writing guide
Machine-checked rules for entry text, applied to entries and drafts.
Type facets
Controlled kind and qualifiers for every entry type label.
Editorial governance (draft)
Roles, review separation, reserved decisions and conflicts of interest.
Source-checking worksheet
One row per open research question, with columns for a researcher to record a checked citation.
Spanish review worksheet
Every Spanish pilot name with its English source, for a native speaker to confirm, correct or reject.
Language pack worksheet
Every smell, descriptor, place and material with its English name and context, for proposing names in a new language. Checked offline; nothing is submitted.
Dataset licence notice
Current licence status of SmellBook-created datasets, generated from one switch.
Dataset release history
Date, digest, node and edge counts, chain hash, notes and archive address of each knowledge-graph dataset release.
Graph entities table
Every knowledge-graph entity as a spreadsheet row, with localized names and review status.
Graph relations table
Every typed relationship as a spreadsheet row, with provenance.
Names index
Every SmellBook name in every language mapped to permanent identifiers, with shared names flagged, for reconciling other datasets.
SmellBook Learn data
Every Learn course, lesson, word pair and classroom guide, linked to the graph by permanent identifier.
Data package
The knowledge-graph dataset as a Frictionless Data Package with digests and table schemas.
DCAT-AP catalogue
SmellBook's public datasets described in the European DCAT application profile used by open-data portals.
Graph health
Orphans, dangling edges, domain and range checks and degree by entity type.
Relation exports
Every edge of one relation type as NDJSON: /data/graph/relations/{type}.ndjson.
Schema registry
JSON Schema definitions and compatibility rules for public records.
Data inventory
Formats, media types, schema mappings and record counts.
Citation metadata
Suggested citation, version and attribution boundaries.
Developer contract
Read-only guarantees, versioning rules and public-surface exclusions.
Route index
Documentation and machine-readable public routes.
Data README
Plain-text starting point for the public data surface.
Version
Milestone, canonical website and data schema version.
Release manifest
Milestone, canonical website, counts, critical routes and machine surfaces.
Checksums
SHA-256 digests for core deterministic public export structures.
idslugnamealiasessummarytypecategorydescriptorsstatuscreatedAtupdatedAtrevisioncanonicalUrlverifiedClaimIdsgraphWeightssimilarityScoresrankingWeightsrecommendationFormulapersonalizationStateuserProfileprivateRelationshipsPublicly readable does not mean unrestricted ownership transfer.
These endpoints make selected public facts and SmellBook-authored catalog fields easier to read by software. They do not grant a blanket license to SmellBook branding, original editorial writing, taxonomy presentation, internal systems, or third-party source material. Individual source rights and licenses remain their own.
Schema version 1.0 is an early public contract. Fields may evolve in later versions, but SmellBook should change the schema version when a breaking change alters the meaning or availability of exported fields.