Questions worth asking

Goable is a physics-first suitability-scoring API for outdoor activities. One REST call to POST /v1/score turns weather, ocean, atmospheric and astronomical inputs into a single 0 to 100 verdict for a specific activity, place and time, with a confidence value and hard-gate context. The scoring path is deterministic and audit-logged; the activity catalogue is open under CC BY 4.0 while the engine is proprietary. See the science reference.

Goable supports the activity profiles published in the versioned open catalogue, for example kitesurfing, surfing, ski touring, paragliding, trekking and road cycling. Coverage varies by activity and location, and the API returns the applicable profile and version together with any coverage or confidence limitations. Browse the open catalogue.

It means the core physical components of a score are grounded in documented scientific methods and sources, for example established wind, wave, thermal-comfort and lightning-risk models. The operational thresholds, provider handling, confidence logic and calibration are versioned and documented. Any machine-learning components, where enabled for calibration or personalisation, are kept outside the core deterministic physical scoring path. See the science reference.

Goable records decision provenance for every score and, where permitted and data is available, evaluates forecast components against documented observations and outcomes. Any public verification release identifies its coverage, reference data, sample size, metrics and limitations, and is not a universal guarantee that a given suitability score is accurate. Brier and CRPS measure forecast quality, not that the suitability verdict itself is correct. See forecast verification.

Provider availability and fallback depend on the activity, geography, horizon and plan. When an eligible source fails, Goable may return a degraded-confidence result, a partial response, or an explicit unavailable status, and the response identifies source freshness and the limitations that applied. It does not silently substitute a guess. See provider skill scoring.

Every forecast-mode score carries a confidence value that is a deterministic fold of components such as forecast skill, provider agreement, profile maturity and data coverage. A low number is a signal to treat the verdict cautiously or to require human review, not a defect to hide. See forecast verification.

No. Goable produces a documented suitability signal with a confidence value and hard-gate context; it is decision support, not a safety system or an automated decision-maker. The customer and qualified local operators retain operational, safety and regulatory decision authority, and Goable does not replace an avalanche bulletin, a coastguard advisory or a qualified instructor's judgement. See hard gates.

Not without the accompanying context, and never as a safety certification or guarantee. Under the acceptable use policy you must keep the maturity, confidence, hard-gate and sample-size caveats attached to any score you surface, and you remain responsible for the accuracy and legal compliance of any claim you make. See the acceptable use policy.

Goable is designed to minimise personal data. For personalisation the customer submits a tenant-generated pseudonymous reference rather than a direct identifier, and Goable processes it as a processor where it remains personal data under GDPR Recital 26. Customers must not submit direct identifiers or special-category data. See the privacy policy.

Primary processing is in the EU: API and compute run in AWS eu-west-1 (Ireland) and the primary Neon Postgres database is in eu-central-1 (Frankfurt), with documented international transfers covered by Standard Contractual Clauses. The full sub-processor list, hosting regions and transfer safeguards are set out in the Privacy Policy, DPA and Security page. See security.

The open catalogue is the set of activity definitions, citations, calibration metadata and review status, published as the @goable-io/profiles-catalog package under CC BY 4.0. Anyone can inspect the curve scoring a spot, propose regional variants, or fork it for research; the scoring engine itself stays proprietary. Browse the catalogue.

Yes. There is an official TypeScript client (@goable-io/sdk, on npm, MIT-licensed, zero-dependency, isomorphic Node and browser) and a Python client (goable-sdk, on PyPI, Python 3.10+). Both are generated from the versioned OpenAPI contract and released alongside supported API versions, with compatibility covered by contract tests and versioning; any language can generate a client from the same OpenAPI 3.1 spec. See the SDK docs.

Yes. Goable publishes its current methodology and synthetic artefacts and runs planned validation streams covering suitability calibration, forecast verification, equipment transition and the Inverse Suitability skill-versus-difficulty work; a drift-events stream is proposed, not yet open. Any public dataset release is subject to consent, sample-size and disclosure-control review, and the k of at least 10 contributors, 1 km grid and 90-day lag are release criteria rather than automatic anonymisation. See the research programme.

Yes. A free developer tier is available subject to the current published quotas, intended for prototyping and integration testing. Quotas, paid plans and overage are listed on the pricing section, which is the canonical source; this page does not restate the numbers so they cannot drift. See pricing.

For product and integration support, email support@goable.io; for security disclosures, email security@goable.io; for legal and data-protection matters, email contact@fabio-carucci.com. Formal security questionnaires and partnership enquiries route through the partnerships desk. See the contact page.