# Regress — Design Notes (v0) A location-based capture game in the spirit of *Ingress*, running as a nostr-native app on Archipelago. Instead of Niantic's fictional "portals," Regress uses real Bitcoin-accepting merchant locations pulled from [BTC Map](https://btcmap.org) as the capturable objects. First scope: **Madeira, Portugal**. --- ## 1. What Ingress actually does (researched) - **Portals** are real-world landmarks agents physically travel to and interact with via the app. In Regress, a **BTC Map place** (merchant/ATM/etc.) is the portal-equivalent. - **XM (Exotic Matter)** is the fictional currency/fuel players collect by walking around; spent on all actions. Needs a Regress equivalent (see open questions below — sats? a synthetic in-app resource? nothing at all?). - **Capturing a neutral portal**: deploying a **resonator** claims it instantly for your faction. Up to 8 resonators per portal ("fully resonated"). - **Mods**: optional items that buff a portal's defense/utility. - **Keys & Links**: hacking a portal yields a *key* to it. That key, used at a *different* fully-resonated friendly portal, creates a **link** between the two — you cannot link a portal to itself. - **Control fields**: three mutually-linked portals form a triangle that "fills in" — this is the actual territory-capture unit, not the portals themselves. - **Factions**: Enlightened (green) vs. Resistance (blue); an AI-controlled third faction, Machina (red), exists in later content. - **Attack/defense**: XMP bursters and Ultra Strikes destroy enemy resonators; a portal with zero resonators reverts to neutral and can be recaptured. - **Anomalies**: live, time-boxed global events tied to real Ingress storyline/lore — probably out of scope for v0. Sources: [ingress.com](https://ingress.com/en), [Fev Games Ingress guide](https://fevgames.net/ingress/ingress-guide/ingress-overview/), [Fev Games glossary](https://fevgames.net/ingress/ingress-guide/glossary/) ## 2. BTC Map as the location data source (researched, live-tested) BTC Map (`teambtcmap/btcmap-api`) is a real, actively maintained, crowdsourced map of Bitcoin-accepting places, backed by OpenStreetMap tag data. It has a public, unauthenticated **REST API v4** at `https://api.btcmap.org/`. Key endpoints we'd actually use: - `GET /v4/places?fields=...` — chronological sync of every place, with `updated_since`/`limit` for incremental sync (bring-your-own cache pattern, good fit for a self-hosted Archipelago app that wants an offline-resilient local copy). - `GET /v4/places/search/?lat=&lon=&radius_km=` — places within a radius. This is what a "load the map around me" or "load this island" view would use. - `GET /v4/places/{id}` — single place detail (name, address, phone, website, socials, `osm_id`/`osm_url`, `verified_at`, icon). - `GET /v4/places/{id}/comments`, `/v4/places/{id}/areas` — community comments and country/community area membership. - `GET /v4/areas?lat=&lon=` — which named areas (country/community) contain a point; `GET /v4/areas/{id|alias}` — area metadata + icon image. No API key needed for reads. Write/RPC endpoints exist for community editing but aren't needed for a read-only game client. **Live-tested against Madeira right now:** a 45 km radius search centered on Funchal (32.7607, -16.9595) returns **167 real, currently-tracked places** — comfortably covers the whole island (~57 km × 22 km). Category breakdown from the actual data: | Count | Category (icon) | |---|---| | 35 | restaurant | | 14 | café | | 13 | storefront (general retail) | | 10 | pharmacy | | 7 | hotel | | 6 | bakery | | 5 each | fitness, hair/barber, "business", medical | | ... | long tail: bars, dentists, spas, tour operators, ice cream, a boat charter, a surf shop, a music venue, a bike rental, etc. | This is real, not a mock — it's what the live API returns today. Full raw dump saved at `data/madeira-btcmap-2026-08-04.json` for reference (167 records: id, lat, lon, name, icon, osm_id, verified_at). Source: [btcmap-api docs](https://github.com/teambtcmap/btcmap-api/tree/master/docs/rest/v4) ## 3. Proposed Regress concept (draft — needs your input) Every BTC Map place in scope starts **unclaimed** (Ingress's "neutral grey"). Players claim it, link claimed places together, and linked triangles form a **control field** over that territory — same core loop as Ingress, but the portals are real Bitcoin-accepting businesses instead of public art/landmarks. ### Open questions before I design further 1. **What does "claiming" cost/require?** Ingress ties everything to XM gathered by physically walking near portals. Options for Regress: - **Physical presence required** (GPS check-in via phone, like Ingress) — most faithful to the genre, but means players must actually be in Madeira. - **Sats/Lightning-gated** — claiming costs a small zap/payment (fits the Bitcoin theme hard, ties into podsteadr's existing Cashu/Lightning plumbing you already have running). - **Nostr-native "proof of visit"** — e.g. NFC/QR at the merchant, or a merchant-signed nostr event, so it's cryptographically real rather than GPS-spoofable. - Some combination. 2. **Factions or free-for-all?** Ingress's whole metagame is two-faction conflict. Do you want teams, or is this more like an individual/community "claim businesses you actually support" thing? 3. **Does "unclaimed" ever mean anything to the real business owner?** i.e. is there a path for the actual merchant to claim their own listing (like podsteadr's owner-exemption pattern), or is Regress entirely player-vs- player over data about businesses that don't participate? 4. **Archipelago integration shape** — same fork as podsteadr had: does Regress run standalone on its own domain (external, identity-aware via the `nostr-provider.js` shim like we just wired for podsteadr), or does it need full orchestrator-managed packaging? I'd default to the podsteadr pattern (external + identity-aware) unless you want it embedded differently. 5. **Nostr event shape for claims/links?** Worth defining a custom kind (or piggybacking parameterized replaceable events) early since that's the actual interop surface other clients could read. ## 4. Next steps once the above is answered - Pull the full Madeira dataset via `/v4/places` (chronological sync endpoint) rather than the radius search, so we get a clean incremental-sync foundation instead of a one-off snapshot. - Define the Regress data model: place (from BTC Map, cached locally) → claim → link → field, plus whatever resource economy answers question 1. - Stand up the repo skeleton (backend + frontend, matching the podsteadr stack unless you want something different) and wire BTC Map sync as the first real feature.