evidoria

← Back to browse

Good practice Imported

Chimple — Offline AI-Adaptive Literacy & Numeracy App Reaches Out-of-School Children Across India, Tanzania and Sierra Leone

India · Bangalore · See the India profile

Evidence: Observational / pre–post Top 8% 73/100 · Ask Evidence Copilot about this practice

A free, offline AI-adaptive literacy/numeracy app for ages 3–8, tested with out-of-school children in Tanzania. An independent Central Square Foundation study found a significant +0.24 SD Math gain per extra hour used, but no significant English gain.

+0.24 SD
Math score gain per additional hour of Chimple use (2024)
+0.36 SD
Math score gain per additional hour of use, first-graders (2024)
+0.13 SD
English score gain per additional hour of use (not statistically significant) (2024)
~17 %
Literacy gain reported across the Tanzania field trial (all five finalist apps) (2017–2019)
~18.2 %
Numeracy gain reported across the Tanzania field trial (all five finalist apps) (2017–2019)
Chimple — Offline AI-Adaptive Literacy & Numeracy App Reaches Out-of-School Children Across India, Tanzania and Sierra Leone

Details

Maturity
Scaling
Promoter
Chimple Learning
Period
2015–present (Tanzania field trial 2017–2019; India dose-response study 2024)
Keywords
early childhood & primary foundational literacy, numeracy, edtech, offline/low-resource learning

Context

Chimple is a free, offline-capable, gamified tablet app built by a Bangalore-based nonprofit to teach foundational literacy and numeracy to children aged 3–8 through more than 60 games and 70 stories in multiple languages. It was one of five finalists — out of 198 teams from 40 countries — in the $15M Global Learning XPRIZE Global Learning Challenge.

Objectives

The app aims to reach out-of-school and low-resource children with adaptive, child-directed literacy and numeracy practice that works without an internet connection.

Activities

UNESCO supported a 15-month field trial across roughly 170 villages in Tanzania in which around 2,000–3,000 out-of-school children aged 7–11 used all five XPRIZE finalist apps interchangeably. Chimple is now distributed through India's government-aligned NIPUN Bharat foundational-learning mission in partnership with the Bharti Airtel Foundation, which launched a stratified randomized evaluation in Haryana in 2022, and has been independently evaluated in low-resource schools in Sierra Leone.

Results

An independent 2024 dose-response study by the Central Square Foundation, run through the LiftEd EdTech Accelerator, found that each additional hour of Chimple use was associated with a statistically significant +0.24 SD gain in Math scores (+0.36 SD for first-graders); the associated +0.13 SD gain in English scores was not statistically significant. Before the Tanzania trial, over 90% of children could not read a single word of Swahili, a share that was roughly halved afterward; Chimple's own figures cite about 17% literacy and 18.2% numeracy gains from that trial, though this result covers all five finalist apps together, not Chimple in isolation.

Conclusions

The evidence is genuinely mixed: Chimple shows a statistically significant, independently measured Math effect, but no significant English effect, and its strongest population-level Tanzania figures cannot be isolated to Chimple alone. In Sierra Leone it was reported to improve attendance and engagement, though without a published effect size, and no current public child-data-protection or algorithmic-transparency policy for the app was found.

Implementation
Indicative cost
Low (< €50k)
Time to results
Long (> 3 years)
Staffing & skills
Bangalore-based nonprofit development team, UNESCO field-trial coordination staff in Tanzania, Bharti Airtel Foundation partnership staff for the India rollout

Conditions for success

  • Offline-capable design suited to low-connectivity, out-of-school settings
  • Multi-language content (60+ games, 70+ stories)
  • Government-aligned distribution channel (NIPUN Bharat) for scale

Common failure modes

  • Effect cannot be isolated when children switch between multiple apps, as in the Tanzania trial
  • No statistically significant English-language gain found in the dose-response study
  • No current public child-data-protection or algorithmic-transparency policy found

Where it fits

Governance type
national mission-aligned nonprofit deployment
Scale
multi-country (India, Tanzania, Sierra Leone)
Income level
low-resource / out-of-school populations

Commonly funded by

National / regional programmes Philanthropic / foundation funding

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Replication kit

Reusable artefacts from this practice — as published by their sources.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful