Public data / Version 1
Results, with their evidence.
Query published computational results. Keep the physical size, reported resources, source dates, and limits together.
Public and read-only. No API key. Size units stay separate, and unreported resources remain null. A sampling benchmark does not establish a chemistry advantage.
Explore the API
Dataset metadata ↗JSON response
Selected result
Endpoints
GET /v1/resultsFiltered, sorted records with total count, limit, and offset.
GET /v1/results/{id}One complete record, including sources and links to related results.
GET /v1/metaRecord counts by approach, workload, unit, role, method, comparison group, and source year.
GET /healthAvailability, schema version, dataset date, and record count.
All routes support HEAD and OPTIONS. CORS allows public reads from any origin. Queries combine with AND; each parameter appears once. Invalid or unknown parameters return 400. Missing records return 404.
Query rules & interpretation
Search is a case- and accent-insensitive substring. Years refer to each record’s documented source date. Partial dates sort at the beginning of their known period without changing their stored precision. Size sorting requires a unit and uses the largest reported size in that unit per record. Ties use ascending record ID.
Page size defaults to 50 in direct API requests and is capped at 200. Offset is capped at 100,000. Search is capped at 200 characters; query strings at 4,096 bytes. Facets count records, once per unit or method. Size ranges do not imply equivalent accuracy or comparable costs.
Example requests
curl 'https://api.cuanti.ca/v1/results?approach=classical&unit=spins&sort=-size&limit=10' curl 'https://api.cuanti.ca/v1/results?claim_role=classical-response&sort=date' curl 'https://api.cuanti.ca/v1/meta'
See the OpenAPI document for the full response schema. Scripts should send a descriptive User-Agent; the hosting edge rejects the generic Python-urllib agent.
import json, urllib.request
request = urllib.request.Request(
"https://api.cuanti.ca/v1/results?method=dmrg",
headers={"User-Agent": "CuanticaResearch/1.0"})
with urllib.request.urlopen(request) as response:
data = json.load(response)