Skip to content

MPowerUP

MPowerUP is a peer-to-peer mutual-aid mobile app for vulnerable populations — people navigating recovery, reentry, houselessness, and poverty. It organises users into trusted Circles: small, end-to-end-encrypted groups where members share safety information, post and respond to help requests, and vouch for one another. Identity is a cryptographic keypair (did:key) held on-device, not an account on someone else's server.

Status: Active — Phase 1 complete, Phase 2 next. Repo: Big-Nerd-Idea/MPowerUp

  • Two design commitments to read first

These two points override anything the older research notes imply, so they are stated up front. - ### The economic model is grassroots mutual aid, not a token economy

MPowerUP's economic direction is mutual aid, time banking, and other non-token community-pooling alternatives — recognising peer care without pricing it. The earlier per-act token (MPWR) direction has been dropped from the live product. The foregrounded direction is described in Grassroots Economy, Time Banking & Mutual-Aid Alternatives. [HYPOTHESIS] — the specific model is not yet decided or piloted; the strongest design hypothesis the team has converged on is a time-banking peer layer paired with a categorically separate stewardship-stipend layer for paid work on shared community commons. - ### The P2P communication stack is a research POC, not a committed architecture

The WebRTC + libp2p + Yjs work that ships in the Phase 1 app is exploratory — proof-of-concept code, not a settled product architecture. The current direction is to review and consider integrating existing open-source tools (Jami was the original anchor; the 2026 reconsideration weighs Briar, SimpleX, Cwtch, and others) rather than committing to a fully custom-built P2P stack. See Technologies for the framing and the comms-architecture survey for the full build-vs-adopt analysis.

  • Design properties

  • Offline-first. Designed to work under unreliable grid power and limited connectivity as a deliberate property, not a fallback — aligned with the BNI Sustainability directive.
  • End-to-end encrypted. Keys live in on-device secure storage; the relay (where one is used) is content-blind.
  • Keypair identity, not accounts. No phone number or email required to participate.
  • Battery-conscious. No polling, no persistent sockets, no continuous GPS without opt-in.
  • Safety-first for the threat model that actually applies. For this population the most likely attacker is often someone the user already trusts (partner, trafficker, institution). See Security.

  • In this section

Page What it covers
Guardian AI The protective-AI design principle
Operations How Circles and the mutual-aid flow work; the grassroots economic model
Technologies Comms-architecture research POC, ADRs, security
Software The app stack (React Native, WatermelonDB, Yjs, libp2p WebRTC)
Hardware Device requirements, offline-first implications, refurb sourcing
- ## Cross-project notes
- rlivn ↔ mpowerup: RlivN's backend may eventually leverage MPowerUP's P2P/CRDT primitives for offline-resilient caregiver sync. Future integration point.
- Second Boot ↔ mpowerup: Refurbished laptops can ship with MPowerUP preconfigured once a desktop/web build exists (Phase 5). See Boot Up.