Curiosity
The Psychedex knowledge graph. Structured pharmacological data, subjective effect mappings, and harm reduction content: versioned, cited, and queryable.
the data
how data gets in
Psychedex data flows through a multi-stage pipeline designed for accuracy and accountability. Source material is extracted from published literature and pharmacological databases, then structured by automated agents. Human editors review, edit, and approve all content before publication. Nothing reaches the site without editorial sign-off.
Every safety-critical value — dose ranges, interaction severities, toxicity assessments — requires two independent citations. Values inferred by AI are labeled as inferred in the database. Uncertainty is surfaced, never hidden.
what's inside
Psychedex organizes knowledge into four parallel content indexes. Substances cover material entities. Practices cover immaterial entities. Effects link them both. Frameworks provides the models and theories that contextualize conscious experience. The effects taxonomy is the common language that connects everything.
curiosity API
Curiosity is the general-purpose data API. It exposes published substance and effect records as structured JSON — dose tiers by route of administration, interactions with severity, duration, tolerance, chemistry. These 7 endpoints serve production traffic today; each one requires an API key.
a request
GET /api/v1/curiosity/substances?page=1&limit=50 Authorization: Bearer <your-key>
every response
{
"data": { ... },
"meta": {
"version": "1.0",
"generated": "2026-08-13T09:41:07.201Z"
}
}the platform
Eight APIs are named. Two of them exist. The badge on each card says which, and the line under the description says what stands behind the name today — including the two whose capability already ships under a Curiosity path, and the four that are a reserved name and nothing more.
Exposes published substance and effect records as structured JSON — dosage by route of administration, interactions with severity, duration, tolerance, chemistry.
Seven endpoints, key-gated.
Returns scheduling and legal status for a substance in a named jurisdiction, plus the jurisdictions that have entries.
One endpoint, key-gated. Jurisdiction coverage is thin — most substances answer 'unknown' for most jurisdictions.
Checks interactions across a set of substances with severity ratings and mechanisms.
No Nexus route. Pairwise checking across up to 10 substances ships today at /api/v1/curiosity/interactions/check.
Cross-modal queries across the effects index.
No Prism route. Effect lookup and effect-to-substance links ship today under /api/v1/curiosity/effects.
Resolves common names, brand names, and chemical identifiers to canonical substance records.
No route. Slug and PSX ID are the only accepted identifiers today.
Data lineage, editorial history, and version diffs for any content entity.
No route. Provenance and version history are published as pages instead: /substances/<slug>/evidence and /substances/<slug>/history.
Aggregates drug-checking results and alert data for contamination and adulteration warnings.
No route and no data sources contracted. See /sentinel for the design and scope.
Substance-specific emergency guidance, symptom triage, and first-responder protocols.
No route. Emergency guidance is published as pages at /harm-reduction/emergency.
principles
status
The 8 endpoints above are live and stable. Getting in is the part that is not: keys are provisioned manually while the program is in private preview. There is no signup form and no key dashboard — write to contact@psychedex.org. Substance data continues to be reviewed and published behind them.