Sources · Every number, with the public reference behind it Updated 2026

If you can’t
fact-check it,
it’s marketing.

Rule 02 of the voice playbook. We try to live by it. Every number printed on this site has a public source. They are listed below, with links and dates, in the order they appear. If you find a number on the site that isn’t here, write to contact@ebstar.co and we’ll either cite it or correct it.

01 · Imali / South Africa§ 01

The SA fintech numbers.

Population sizing for SA’s informal economy. We use the FinScope figure because it is the survey that asks the question we care about: businesses banks can’t see.

1.7 million
SA informal businesses (56% of 3.0M total MSMEs). FinMark Trust, FinScope MSME South Africa 2024 wave (3,027,298 MSME owners operating 3,912,277 businesses; informal share 56%). Headline figure aligned with StatsSA QLFS. finmark.org.za.
R2.3T informal economy
Annual gross of the SA informal sector. SA Reserve Bank / TIPS estimates referenced in the Imali fundraise intelligence pack.
$331B credit gap
Sub-Saharan Africa MSME credit gap. IFC, MSME Banking Handbook 2017 to 2018 methodology, confirmed in the IFC 2025 update.
SARS TT03
Small-Business Tax Rebate / TT03. South African Revenue Service tax type for small-business income. Imali’s SARS-aligned categories use the TT03 schema. Filing in queue.
Turnover Tax 2026
R600K tax-free, R2.3M ceiling. SARS Turnover Tax thresholds, effective 1 April 2026. The compliance hook for Imali’s tax module.
POPIA Art 72
Cross-border data transfer. Protection of Personal Information Act, 2013, Section 72. The article that governs Imali’s standard contractual clauses for the EU-region backend. popia.co.za / Section 72.
POPIA Art 24
Right to deletion / correction. Section 24 governs the data subject’s right to request correction or destruction. The basis for Imali’s delete me reply flow.
02 · Chingu / Korea§ 02

The Korean eldercare numbers.

Population sizing and isolation indicators come from Statistics Korea (KOSIS) and the Ministry of Health and Welfare (보건복지부 / MOHW). The relevant statutes are PIPA and the AI Basic Act.

2.13 million
Korean seniors living alone. Statistics Korea (통계청 / KOSIS), Population & Housing Census 1인가구 (one-person household) tables, “65세 이상 1인가구.” Latest published year referenced. kosis.kr.
32.6%
Seniors with no daily conversation partner. 보건복지부 / Ministry of Health and Welfare, 노인실태조사 (National Survey of Older Koreans). The MOHW publishes this triannually; the figure is cited from the most recent wave. mohw.go.kr.
3,924 (2024)
Solitary deaths recorded in Korea, 2024. 보건복지부, 고독사 실태조사 (Lonely Deaths Survey). The official annual count under the “Act on the Prevention and Management of Lonely Deaths.” mohw.go.kr.
PIPA Art 23
Sensitive information / biometrics. Personal Information Protection Act, Article 23: biometric data (including voiceprint) requires separate, opt-in consent. The basis for Chingu’s separate voice-consent flow. pipc.go.kr.
PIPA Art 21
Destruction of personal information. The article behind Chingu’s 90-day auto-delete on voice clips and transcripts.
AI Basic Act Art 31
AI identity disclosure. 인공지능 발전과 신뢰 기반 조성 등에 관한 기본법 / AI Framework Act. Article 31 requires AI services to disclose their AI nature on initiation. The basis for Chingu’s first-call disclosure script.
지역사회통합돌봄법
Integrated Community Care Assistance Act. Effective March 2026. Legally obligates Korean municipalities to provide regular welfare monitoring for elderly residents living alone. The demand-side push for KONEPS procurement of products like Chingu.
Super-aged threshold
20%+ population aged 65+. Korea crossed this in December 2024, a year ahead of projection. Source: 통계청 / Statistics Korea, 2024 announcement.
KONEPS / 나라장터
Korean government procurement platform. All municipal contracts pass through KONEPS. Vendor registration is the post-seed gate for paid pilots.
KISA
Korea Internet & Security Agency. Information Security Management certification required for higher-value municipal contracts. Chingu cert target: Q3 2026.
03 · Product numbers§ 03

Numbers we measure ourselves.

Performance and pricing claims are measured internally and listed here so a journalist or partner can check them against the live products.

700ms
Median Chingu voice-AI response time. Measured on the staging build, p50 across Standard Korean and Gyeongsang corpora. p90 published on launch.
60s
Imali first-message-to-logging time. Measured against pilot WhatsApp threads. Median across the first six weeks of pilot.
R0 / R49 / R99
Imali pricing tiers, ZAR / month. Published in product. Free tier capped at 20 transactions / month.
₩10,000 / resident / month
Chingu municipal pricing. Anchor price for municipal procurement; included scope (calls / day, alert SLA, dashboard seats, K-language support hours) is named on the procurement landing page.
₩12,900 / month
Chingu family-tier pricing. First month free.
5M+ streams
Ebstar music streams across platforms. Aggregated from Spotify for Artists, Apple Music for Artists, YouTube Analytics. Independent distribution. Cited on the founder page as a distribution credibility marker, not as engineering proof.
04 · Composite scenes§ 04

What composite means here.

Some scenes on the journal are composites: single narrative voices that braid details from several real conversations. We use this format when participants have asked to remain unnamed during pilot. We label every composite inline, at the top of the piece, never just in a footer.

Composites are not testimonials. Direct quotes inside a composite do not represent a single named person. Numbers (transactions logged, ages, dates) are factual where possible and noted otherwise.

When a participant opts in to be named, the composite is replaced with an attributed scene. Until then, the composite tag stays on.

Read the voice playbook Rule 04 · Honest about stage →

Spot a number that isn’t here?

Either we missed it, or we shipped a sentence that fails our own rule. Both are worth a note. Write to contact@ebstar.co with the URL and the line. We’ll either cite it within a week, or correct it.

contact@ebstar.co The five-question test →