|
| 1 | +# Supabase — Phase 2B Classification |
| 2 | + |
| 3 | +Status: CLOSED (committed) |
| 4 | +Phase A re-read during B: NO (isolation rule honored — answers "what do I see", not "is the prediction confirmed") |
| 5 | +Evidence opened: YES (Phase A locked + pushed in commit 22eaafd before this file) |
| 6 | +Classification author: Claude (operator-reviewed) |
| 7 | +Created: 2026-06-06T15:23:27+0000 |
| 8 | + |
| 9 | +> Regla de aislamiento (PHASE2_PROTOCOL § Fase B): la clasificación responde "¿qué veo en los datos?", no "¿se confirma mi predicción?". La predicción LP no participa de la elección de Primary Strategy. |
| 10 | +
|
| 11 | +--- |
| 12 | + |
| 13 | +## 0. Anchor data (evidence) |
| 14 | + |
| 15 | +- **Founded:** 2020, by Paul Copplestone & Ant Wilson (Y Combinator). Positioned as "the open-source Firebase alternative". |
| 16 | +- **License (root, master HEAD):** Apache License 2.0 (primary source: raw.githubusercontent.com/supabase/supabase/master/LICENSE, copyright "2024 Supabase", fetched 2026-06-06). No relicensing event found; explicit negative search for BSL/SSPL relicensing returned nothing on Supabase itself. Counter-signal: Supabase RE-licensed an acquisition (Logflare, previously BSL) TO Apache 2.0 — the opposite direction of a licensing pivot. |
| 17 | + |
| 18 | +--- |
| 19 | + |
| 20 | +## 1. Trajectory |
| 21 | + |
| 22 | +**Value:** `native` |
| 23 | + |
| 24 | +Evidence: open-source core (Apache 2.0) and a hosted commercial platform both existed from the 2020 launch; the repo shipped self-host tutorials from day 1. No relicensing event and no posture pivot — the current state is the original one. |
| 25 | + |
| 26 | +--- |
| 27 | + |
| 28 | +## 2. Distribution Posture |
| 29 | + |
| 30 | +**Value:** `cloud-first` |
| 31 | + |
| 32 | +Evidence: both routes are production-grade — self-host is documented as "production-parity" with the hosted platform (11-service docker-compose stack) — but the managed cloud is the recommended/promoted route (default "Start your project" CTA, free tier, professional support; self-host is community-supported and lacks managed-only features such as branching, PITR, managed backups, advanced metrics, analytics buckets, ETL). |
| 33 | + |
| 34 | +--- |
| 35 | + |
| 36 | +## 3. Primary Strategy |
| 37 | + |
| 38 | +**Value:** `DND` (Dual-Native Design) — estado: **CONFIDENT** (with a documented mild tension toward CP; see § 5) |
| 39 | + |
| 40 | +Evidence: |
| 41 | +- The Apache-2.0 core (Postgres, auth, storage, realtime, edge functions) is fully self-hostable at production parity, including full `postgres` superuser access that the managed platform deliberately restricts. Self-host is NOT deprecated or degraded. |
| 42 | +- Monetization arises from managed **operational convenience** — managed backups/PITR, autoscaling, branching, advanced metrics, compliance/support — not from deliberate degradation of self-host nor from a license restriction. This matches DND inclusion ("monetización surge de conveniencia operativa, no de degradación deliberada del self-host") and explicitly fails the CP inclusion test ("self-host deprecado o desalentado"). |
| 43 | + |
| 44 | +--- |
| 45 | + |
| 46 | +## 4. Secondary Strategies |
| 47 | + |
| 48 | +**Value:** none registered. |
| 49 | + |
| 50 | +Rationale: new platform capabilities ship cloud-first (a faint Commercialization-Pivot flavor), but the production-grade core remains fully open and self-hostable, so this does not rise to a Secondary Strategy. It is recorded as a tension in § 5 rather than promoted to a category (Regla 4: a strategy is born only when classification fails honestly, which it did not here). |
| 51 | + |
| 52 | +--- |
| 53 | + |
| 54 | +## 5. Taxonomy Stress |
| 55 | + |
| 56 | +**Value:** `ambiguo leve` (mild ambiguity) |
| 57 | + |
| 58 | +Why: the boundary between "managed convenience" (DND) and "self-host degradation via cloud-only feature velocity" (CP) is judgment-dependent. Platform-exclusive features (branching, PITR, ETL, analytics buckets) create real asymmetry between routes — but the core remains self-hostable at production parity, so the case resolves to DND without forcing. Not severe, not UNCLASSIFIABLE. |
| 59 | + |
| 60 | +**Explicit non-stress note (orthogonality working as designed):** Primary Strategy = DND while Distribution Posture = cloud-first is NOT a contradiction. Strategy captures the value-capture mechanism (managed convenience, self-host viable); Posture captures what is promoted today (cloud). Supabase is a clean demonstration of the Strategy ⊥ Posture orthogonality the taxonomy claims — evidence FOR the multidimensional frame, not against it. |
| 61 | + |
| 62 | +--- |
| 63 | + |
| 64 | +## 6. Bookkeeping |
| 65 | + |
| 66 | +- working_classification: **true** (Regla 3 — pre-full-research working hypothesis; does not count for metrics/synthesis/public outputs until the 10-point Definition of Done is met). |
| 67 | +- LP role observed: **absent** (Apache 2.0 throughout; primary-source confirmed). |
| 68 | +- Cohort observed: **post-pattern** (founded 2020 ≥ 2018). |
| 69 | +- timestamp: 2026-06-06T15:23:27+0000 |
0 commit comments