Bundu Foundation Contact

Bundu open standards

Six benchmarks. Built for Africa. Open to the world.

We don't measure against US, EU, or Asian markers. Where existing African standards apply, we adopt them. Where they are missing, the Foundation publishes new open standards that any framework, school, institution, or operator can audit against. Free to adopt, free to improve, attribution to the Bundu Foundation.

Back to research See the frameworks

  • Draft v0.3

    Outdoor-readability

    Type, colour, and layout that hold up at midday on a verandah, not just on a glossy laptop in an office.

    Benchmarks

    • APCA Lc 90+ contrast on body text (not WCAG ratio alone — APCA is the perceptual successor).
    • Body type size and weight that remain legible under direct-sun glare on a typical sub-USD 200 smartphone.
    • Touch targets ≥ 56 px on primary actions.
    • Anti-glare considerations in product copy (avoid pure-white backgrounds with thin grey type).

    Seeded from

    APCA (Accessible Perceptual Contrast Algorithm) — community-led successor to WCAG 2 contrast. We adopt the algorithm and publish African-classroom test conditions.

  • Draft v0.4

    Connectivity-gradient

    A floor, not a target. Products must work down the connectivity gradient — fibre, 4G, 3G, 2G/SMS, offline — without failing silently.

    Benchmarks

    • 3G is the floor; ~77% of African population is covered by 3G (GSMA Mobile Economy Africa 2025).
    • Per-route JavaScript budget ≤ 100 KB gzipped on the critical path.
    • Offline-first writes with queued sync; no toast saying "you are offline" that loses the user's work.
    • Data-bundle-aware media: progressive images, audio over video, captioned video when delivered.
    • SMS / USSD fallback for transactional flows (fee balance lookups, exam-result delivery, sign-in OTPs).

    Seeded from

    A4AI 1-for-2 affordability target; GSMA coverage gap framework. The Foundation publishes the per-route engineering budget.

  • Draft v0.2

    Mother-tongue-first delivery

    Lower-primary content in a learner's first language; transition to the language of instruction is staged, not abrupt.

    Benchmarks

    • Lower-primary (≈ ages 5–8): instruction in L1; content packs in at least one regional language cluster.
    • Upper-primary: parallel L1 + LMI content for transition years.
    • Voice-first interaction where literacy is still emerging.
    • Tooling: Kolibri language packs, African Storybook, Bloom Library, Ubongo, eLimu integration.
    • Mainstream LLM responses must be evaluated for L1 quality; Lelapa InkubaLM and Masakhane models considered first for supported languages.

    Seeded from

    UNESCO mother-tongue-first guidance plus national curriculum policies (Tanzania, Rwanda, Nigeria, South Africa, Ethiopia). The Foundation publishes a single-page benchmark schools can audit against.

  • Draft v0.3

    Mobile-money-native fee surfaces

    Mobile money is the household economy in Africa. Fee surfaces, vendor payouts, and stipends integrate at least one mobile-money rail by default.

    Benchmarks

    • Native integration with at least one of: Daraja (M-Pesa), MoMo Developer API, Airtel Africa API, Orange Developer, EcoCash, Wave, or an aggregator (Flutterwave, Paystack, Cellulant, DPO, MFS Africa, Onafriq).
    • Idempotent fee references — handle USSD timeouts, double-debits, and reversals safely.
    • Reconciliation cadence ≤ 24h with auditable trail.
    • Receipts via WhatsApp / SMS with optional USSD lookup for unconfirmed payments.
    • Fraud-pattern awareness — SIM swap, Wangiri, phishing — surfaced in user UX, not hidden.

    Seeded from

    GSMA mobile-money fraud guidance; provider developer documentation. The Foundation publishes an integration-test checklist any school IT can run before go-live.

  • Draft v0.4

    Off-grid power resilience

    32% of primary and ~50% of secondary African schools are off-grid; in countries with grid, 4–16 hours of load-shedding is routine. Power is a design parameter, not a contingency.

    Benchmarks

    • Classroom power budget: 30-learner classroom on phones + projector ≤ 300 W peak; full computer lab ≤ 2,000 W peak.
    • Solar sizing: 3 kW small primary / 5 kW medium / 6–10 kW secondary with lab. Lithium-iron-phosphate batteries at ≤ USD 100/kWh.
    • Backup hierarchy: solar + battery → grid → genset → UPS for LMS server only.
    • Charging discipline: device-fleet charging during solar peak (10:00–14:00); daytime-bundle data plans aligned to teaching hours.
    • Server consolidation: a single Raspberry Pi 5 (~10 W) serving Kolibri offline content to a school-wide WiFi.

    Seeded from

    European Commission Joint Research Centre, Solar power offers a brighter future for African schools (Feb 2025). The Foundation publishes a sizing worksheet and procurement reference.

  • Draft v0.2

    Multilingual UX baseline

    African UX is multilingual by default. The minimum is not English-plus-French; it is regional language clusters with proper layout, voice, and input support.

    Benchmarks

    • East Africa: Swahili, Amharic, Oromo, Luganda, Kinyarwanda. West Africa: Hausa, Yoruba, Igbo, Akan/Twi, Wolof, Fula. Southern Africa: isiZulu, isiXhosa, Sesotho, Shona, Ndebele, Chichewa. North Africa: Arabic + Tamazight + French. Lusophone: Portuguese + Umbundu/Makua. Francophone: French + national languages.
    • Reflowable layouts that absorb 2–3× text expansion without breaking grids.
    • Right-to-left support for Arabic.
    • Voice-first interactions where written literacy is still emerging.
    • Mainstream LLM integrations evaluated for L1 quality before deployment.

    Seeded from

    Country language-of-instruction policies; UNESCO mother-tongue guidance; Masakhane translation benchmarks. The Foundation publishes a per-region minimum-viable language list.

Methodology

How a Bundu open standard ships.

  1. 1. Research identifies the gap.

    A pan-African pattern that existing standards do not capture — outdoor-readability under midday sun, mobile-money fee surfaces, mother-tongue-first delivery, load-shedding as a design parameter — becomes a research note.

  2. 2. Draft a measurable benchmark.

    The Foundation drafts a minimum, testable benchmark — not an aspirational principle. A school IT lead, a product team, or a partner ministry can audit against it without buying a tool or hiring a consultant.

  3. 3. Apply it through a framework.

    The first proof point is a framework that uses the standard in anger. Our three current frameworks — K-12 Digital Campus, Digital Literacy, K-12 Support Process — benchmark against the standards on this page.

  4. 4. Open it to contribution.

    Every standard is open. Partners, schools, and operators adopt, audit, and contribute back. Versions move from v0.x draft to v1.0 stable when multiple independent deployments report against them.

Adopt a standard. Run a benchmark. Send the result back.

The Foundation will publish anonymised benchmark data so the whole continent can see what good looks like in practice — not in theory.

Contribute a benchmark All research