CB1 Token Schema v1.0 -- Open Standard for Cannabis Product Intelligence

Section 1

Introduction

Cannabis is the only regulated consumer product category without a structured compound intelligence standard. Certificates of Analysis -- the documents that define what is in a cannabis product -- exist as unstructured PDFs with no interoperability, no scoring methodology, and no machine-readable format.

The CB1 Token Schema defines a structured, versioned, research-grounded data object called a Product Intelligence Token. This token transforms raw COA data into actionable intelligence: effect predictions, consumer need matching, compound interaction analysis, and multi-audience content -- all anchored to public blockchain infrastructure for provenance verification.

1.1 Scope

This specification defines the structure, data types, required fields, optional fields, compound taxonomy reference, effect scoring methodology, customer need mapping, and provenance chain requirements for a CB1 Product Intelligence Token.

This specification does not define how tokens are generated, how they are stored, or how they are displayed. It defines what a token contains and how that content is structured.

1.2 Terminology

Term	Definition
Token	A Product Intelligence Token: the structured data object defined by this schema.
COA	Certificate of Analysis. The source laboratory document from which a token is derived.
Compound	A cannabinoid or terpene identified in the COA and tracked in the CB1 Compound Taxonomy.
Effect Dimension	One of 20 scored dimensions representing a predicted physiological or psychological effect.
Customer Need	One of 20 consumer use-case categories derived from weighted effect dimension scores.
Provenance	Cryptographic proof that a token's immutable data has not been modified since creation.
Anchor	The act of recording a cryptographic hash of token data on a blockchain.

1.3 Design Principles

#	Principle	Description
1	Lab-verified data only	Tokens are derived from licensed laboratory COAs, not user-submitted data.
2	Deterministic scoring	Given identical compound inputs, the scoring engine must produce identical outputs. No LLM in the scoring loop.
3	Immutable provenance	The COA hash and compound profile hash are anchored to a public ledger at creation and cannot be retroactively modified.
4	Mutable intelligence	Effect scores and narratives may be regenerated as the taxonomy evolves. Provenance covers source data, not derived intelligence.
5	Multi-audience	A single tokenization event produces intelligence for laboratories, brands, retailers, and consumers simultaneously.

Section 2

Token Structure

A conforming Product Intelligence Token MUST contain the following top-level objects.

2.1 Root Object

Field	Type	Req	Description
`token_id`	UUID v4	REQUIRED	Globally unique identifier.
`schema_version`	string	REQUIRED	Schema version. Value: "1.0".
`created_at`	ISO 8601	REQUIRED	Timestamp of token creation.
`source`	object	REQUIRED	COA source metadata.
`cannabinoids`	array	REQUIRED	Cannabinoid compound objects.
`terpenes`	array	REQUIRED	Terpene compound objects.
`contaminants`	array	OPTIONAL	Contaminant test results.
`metadata`	object	REQUIRED	Computed summary values.
`intelligence`	object	REQUIRED	Effect scoring and need mapping.
`narratives`	object	OPTIONAL	Multi-audience generated content.
`blockchain_provenance`	object	REQUIRED	Immutable provenance anchors.

2.2 Source Object

Field	Type	Req	Description
`batch_id`	string	REQUIRED	Laboratory batch or sample identifier.
`strain_name`	string	OPTIONAL	Product or strain name from COA.
`sample_type`	enum	REQUIRED	FLOWER, CONCENTRATE, VAPE, EDIBLE, TINCTURE, BEVERAGE, CAPSULE, TOPICAL, PRE_ROLL, INFUSED_PRE_ROLL, LIVE_RESIN, LIVE_ROSIN, DISTILLATE, RSO, HASH.
`lab_name`	string	REQUIRED	Name of the testing laboratory.
`lab_license`	string	OPTIONAL	Laboratory license number.
`state`	string	REQUIRED	US state code (2-letter).
`test_date`	ISO 8601	REQUIRED	Date the COA was issued.

Section 3

Intelligence Output

The intelligence object contains the output of the deterministic scoring engine. All scores are floating-point values between 0.0 and 1.0.

3.1 Effect Profile (20 Dimensions)

A conforming token MUST include an effect_profile object with exactly 20 keys. The scoring methodology is deterministic: given identical compound inputs, the scoring engine MUST produce identical outputs.

Key	Label	Group
`relaxation`	Relaxation	Experience
`euphoria`	Euphoria	Experience
`energy`	Energy	Experience
`focus`	Focus	Experience
`creativity`	Creativity	Experience
`sedation`	Sedation	Experience
`sensory_enhancement`	Sensory Enhancement	Experience
`introspection`	Introspection	Experience
`social_ease`	Social Ease	Experience
`arousal`	Arousal	Experience
`stress_relief`	Stress Relief	Wellness
`mood_elevation`	Mood Elevation	Wellness
`muscle_relaxation`	Muscle Relaxation	Wellness
`anti_inflammation`	Anti-Inflammation	Wellness
`anti_nausea`	Anti-Nausea	Wellness
`pain_relief`	Pain Relief	Therapeutic
`neuroprotection`	Neuroprotection	Therapeutic
`anti_spasm`	Anti-Spasm	Therapeutic
`appetite_stimulation`	Appetite Stimulation	Therapeutic
`appetite_suppression`	Appetite Suppression	Therapeutic

3.2 Customer Need Map (20 Categories)

A conforming token MUST include a customer_need_map object with exactly 20 keys. Each need score is derived from a weighted combination of effect dimension scores.

Key	Label	Primary Effect Drivers
`sleep`	Sleep	sedation, relaxation
`anxiety`	Anxiety Relief	stress_relief, relaxation
`chronic_pain`	Pain Management	pain_relief, anti_inflammation
`recreation`	Recreation	euphoria, social_ease, creativity
`focus_work`	Focus / Productivity	focus, energy
`social`	Social	social_ease, euphoria
`nausea_gi`	Nausea / GI	anti_nausea, appetite_stimulation
`migraine_headache`	Migraine	pain_relief, anti_inflammation
`inflammation_arthritis`	Inflammation	anti_inflammation, pain_relief
`muscle_recovery`	Muscle Recovery	muscle_relaxation, anti_inflammation
`mood_depression`	Mood / Depression	mood_elevation, euphoria
`stress_ptsd`	Stress / PTSD	stress_relief, relaxation
`appetite_management`	Appetite Control	appetite_suppression, energy
`seizure_epilepsy`	Seizure Support	anti_spasm, neuroprotection
`neurological_support`	Neurological	neuroprotection, anti_inflammation
`skin_wellness`	Skin Wellness	anti_inflammation
`cancer_support`	Cancer Support	anti_nausea, pain_relief
`metabolic_health`	Metabolic Health	appetite_suppression, anti_inflammation
`intimacy`	Intimacy	arousal, sensory_enhancement
`meditation`	Meditation	introspection, relaxation

Section 4

Compound Taxonomy Reference

CB1 Compound Taxonomy v2.1 is the scientific foundation of the scoring engine. All compound names in a conforming token MUST use canonical names defined in this taxonomy.

Research Database Extension

Beyond the 41 curated compounds below, the CB1 intelligence layer draws on 6,200+ research compounds from the Cannabis Compound Database (University of Alberta) with 256 compound-protein receptor binding links. Compounds not in the curated taxonomy are dynamically scored via the receptor-effect mapping matrix with a 0.85x inference discount.

4.1 Cannabinoids (11)

Canonical Name	Classification	Primary Receptor Affinities
`thc`	Major	CB1, CB2
`cbd`	Major	5-HT1A, TRPV1, GPR55, PPAR-y
`thca`	Major	TRPA1, TRPM8
`cbda`	Major	5-HT1A, COX-2
`cbg`	Minor	CB1, CB2, 5-HT1A
`cbga`	Minor	PPAR-y
`cbn`	Minor	CB1, CB2, TRPV2
`cbc`	Minor	TRPV1, TRPA1
`thcv`	Minor	CB1, CB2
`delta-8-thc`	Minor	CB1, CB2
`cbdv`	Minor	TRPV1

4.2 Terpenes (30)

Canonical Name	Primary Receptor	Aroma Profile
`myrcene`	TRPV1, GABA-A	earthy, musky, herbal
`limonene`	5-HT1A, A2A	citrus, lemon, orange
`linalool`	GABA-A	floral, lavender
`alpha-pinene`	AChE inhibitor	pine, fresh, forest
`beta-pinene`	AChE inhibitor	pine, woody, herbal
`beta-caryophyllene`	CB2	peppery, spicy, woody
`humulene`	TRPV1	earthy, woody, hoppy
`terpinolene`	5-HT1A	floral, piney, herbal
`ocimene`	--	sweet, herbal, woody
`bisabolol`	TRPM8	floral, chamomile
`nerolidol`	--	woody, floral, citrus
`guaiol`	--	piney, woody, rose
`camphene`	--	camphor, herbal, woody
`geraniol`	--	rose, floral, citrus
`valencene`	--	citrus, sweet, woody

Additional terpenes (15): fenchol, borneol, terpineol, cis-nerolidol, trans-nerolidol, alpha-terpinene, gamma-terpinene, para-cymene, eucalyptol, isopulegol, sabinene, alpha-phellandrene, delta-3-carene, alpha-cedrene, farnesene.

4.3 Receptor Systems (13 Mapped)

The schema maps 13 primary receptor systems to body map positions. The research database tracks 124 unique protein targets across 256 compound-protein binding links from peer-reviewed studies.

Receptor	System
`CB1`	Central Nervous System
`CB2`	Immune and Peripheral
`5-HT1A`	Serotonin
`TRPV1`	Pain / Heat
`GABA-A`	Calming
`GPR55`	Novel Cannabinoid
`PPAR-y`	Metabolic
`AChE`	Memory / Cholinergic
`Adenosine A2A`	Dopamine Modulation
`TRPV2`	Immune Sensing
`TRPA1`	Nociceptive
`TRPM8`	Thermosensory
`COX-2`	Inflammatory Pathway

Section 5

Narrative Content

A conforming token MAY include a narratives object containing generated text content derived from the token's compound data and intelligence output. Narrative content is audience-specific and SHOULD be regenerated when the scoring methodology or research corpus is updated.

Audience	Field	Purpose
Consumer	`quick_learn`	60-second consumer summary of what to expect.
Consumer	`deep_dive`	Research-grounded explanation of compound profile and effects.
Retail	`merch_copy`	Menu description for dispensary product listings.
Retail	`budtender_tips`	Talking points for budtender-customer conversations.
Wholesale	`sales_pitch`	Wholesale pitch positioning product for dispensary buyers.

Section 6

Blockchain Provenance

A conforming token MUST include a blockchain_provenance object documenting the cryptographic anchoring of immutable token data to one or more public distributed ledgers.

6.1 What Is Anchored (Immutable)

SHA-256 hash of the original COA PDF document. SHA-256 hash of the canonical compound profile JSON (sorted by compound name, values truncated to 2 decimal places). Token ID, creating organization ID, and creation timestamp.

6.2 What Is NOT Anchored (Mutable)

Effect scores, customer need scores, and interaction flags. Narrative content. Brand assignments, token status, and consumer-facing metadata. These are intelligence outputs that improve over time -- anchoring them would freeze the product.

6.3 Chain Requirements

Hedera HCS

Scientific Provenance

COA hash, compound profile hash, token ID, org ID, timestamp. Append-only topic. Governed by 39 Fortune 500 council members.

REQUIRED

Ethereum L1

Legal Anchor

Weekly Merkle root of all tokens created in the period. Court-admissible provenance.

OPTIONAL

XRPL

Financial Compliance

XLS-20 NFToken for ownership chain, clawback, and global freeze capabilities.

OPTIONAL

6.4 Verification

Any party with access to a Hedera mirror node or explorer can independently verify that a token's COA and compound profile have not been modified since creation, without relying on CB1's API or infrastructure.

Section 7

Token Lifecycle

A Product Intelligence Token progresses through a governed lifecycle that establishes chain-of-custody from laboratory creation through consumer access.

1

Laboratory

Upload COA

Token minted. Compounds extracted and normalized. Effect scoring runs. Hedera HCS anchor fires. Lab org stamped.

minted

2

Laboratory

Assign to Brand

Brand org ID linked. Brand receives token in their portal inbox.

pending

3

Brand

Accept

Brand confirms the token represents their product.

accepted

4

Brand

Publish

Token visible to retailers and consumers. QR code active.

published

5

Retailer

Display

Token data flows to POS systems, menus, and budtender tools.

6

Consumer

Scan / View

Full intelligence page accessible via QR code or direct link.

Section 8

Conformance Requirements

A system that produces Product Intelligence Tokens conforming to this specification MUST satisfy the following requirements:

Tokens MUST contain all fields marked REQUIRED in this specification.
Compound names MUST use canonical names from CB1 Compound Taxonomy v2.1.
Effect scoring MUST be deterministic: identical compound inputs MUST produce identical effect_profile outputs.
effect_profile MUST contain exactly 20 keys as defined in Section 3.1.
customer_need_map MUST contain exactly 20 keys as defined in Section 3.2.
All effect and need scores MUST be floating-point values between 0.0 and 1.0.
Tokens MUST be anchored to at least one public distributed ledger with a SHA-256 hash of the COA PDF and compound profile JSON.
Token source data MUST originate from a Certificate of Analysis issued by a licensed testing laboratory.
Narrative content MUST NOT make medical claims. All content is educational and informational.
Systems implementing this schema MUST declare the schema_version in every token produced.