Skip to content
JANSETU AI
FROM VOICE TO ACTION
Digital Public Good · India-first

JANSETU AI — Your voice
can shape
your community.

JANSETU AI turns citizen requests into evidence-backed development priorities for governments.

Citizen voice → AI understanding → Evidence fusion → Transparent prioritization → Human decision → Impact. Minimal. Trustworthy. Measurable.

Raise a Community Need
No account to preview · 2 minHow it works
I am acitizen·government agency
Privacy-preservingHuman-governed
What happens next: Speak→ AI understands→ Evidence + score
Demand preview
Vadodara · sample · centroids only
CITIZEN REQUESTS
4,218 clustered
Priority 94 · High
Deterministic v1
Evidence-fused · human-governedSample data · not live
Product · Live preview

See the product, not just the promise.

Citizen PWA and government dashboard — real flows, sample data clearly labeled.

GU · HI · EN
Citizen — Submit
Voice / Text / Photo → AI preview
GU Gujarati voice detected
“અમારા ગામનો રસ્તો…”
Village X, Vadodara · 📷 attached
Preview submission →
Deterministic v1
Government — Priority queue
Evidence · score · audit
94
Vadodara roads
4,218 req · 12.4k · high
62
Water — Surat
892 req · moderate
45 → 28 min
Impact — Baseline → Actual
Observed vs modeled
45
Baseline
22
Target
28
Actual
Observed · verified
Citizen submit →Government dashboard →Docs →
Process · Six steps

Citizen voice → Public action

Not a complaint portal. A public intelligence layer that listens, understands, fuses evidence, and recommends — human always decides. Learn how it works · Developer docs

STEP 01

Citizen voice

A citizen speaks in Gujarati, Hindi or English, types, or photographs the problem. Nothing is required beyond the need itself — location can come from a PIN code or the device.

Raise a community need
Accepted input
Voice · text · photo
Languages on intake
ગુજરાતી · हिन्दी · English

Figures shown are the sample Vadodara cluster used throughout this page — labelled sample, never presented as live.

LISTEN
Multilingual citizen voice
Voice, text, photos — in Gujarati, Hindi, English. Preserve meaning, flag ambiguity.
Evidence-led
UNDERSTAND
AI organizes feedback
Gemini extracts category, urgency, location, affected groups — validated server-side.
Evidence-led
PRIORITIZE
Evidence-backed priorities
Deterministic score: demand 30% + gap 20% + pop 15% + vuln 15% + urgency 10% + feas 10%.
Evidence-led
MEASURE
Track real impact
Baseline → Target → Actual with observed vs modeled, audit-logged.
Evidence-led

JANSETU AI Impact — Measurable outcomes

Impact · Sample data

Measure what changed — not what was promised.

Baseline → Target → Actual with observed vs modeled, labeled clearly. No inflated claims.

0+
Citizen Requests
last 12 months · sample
0
Demand Hotspots
BigQuery GIS · sample
0
Priority Projects
candidate · sample
0
People Impacted
estimated · sample
See measured impact Government dashboardAPI — baselines
TRUST · GOVERNANCE

Why JANSETU AI is trustworthy

Digital Public Good — privacy-preserving, fairness-constrained, human-governed. Frontend is untrusted: all validation, clustering, scoring, recommendations and audit live server-side. Gemini may understand, translate, classify and explain — it must not invent evidence, alter weights silently, approve funding, or override authoritative data. Every priority score persists every component plus weightVersion for reproducibility.

  • Never uses religion, caste, or political affiliation in any scoring decision.
  • Voice is transcribed to text; no biometric profiling, no sales, no ads. Location is opt-in — text locality works.
  • Public analytics show cluster centroids and GeoJSON hotspots — never individual citizen positions.
  • Fully audit-logged: audit logs and system health are transparent.
About JANSETU →Privacy →Accessibility →Contact →
METHOD · DETERMINISTIC V1

Deterministic Priority Engine v1

Priority = demand 30% + infrastructure gap 20% + population 15% + vulnerability 15% + urgency 10% + feasibility 10%. Weights are pinned and versioned — Gemini explains, never overrides.

Demand 30%92
4,218 requests clustered near Vadodara
Gap 20%88
Road index 38/100 vs target 70
Population 15%80
12.4k affected, density-aware
Vulnerability 15%82
SC/ST, women, children weighted
Urgency 10%90
Monsoon closure, 45→28 min
Feasibility 10%64
Terrain, cost, timeline
Band: critical ≥80 · high 65–79 · moderate 45–64 · low <45

See docs and API reference for full spec and OpenAPI.

Scale · BRICS-ready

Built for diverse communities. Designed to scale across borders.

India-first, BRICS-ready — languages, admin hierarchy, currency, datasets per country. Explore BRICS configuration →

🇧🇷 Brazil — Português🇷🇺 Russia — Русский🇮🇳 India — हिन्दी · ગુજરાતી · English🇨🇳 China — 中文🇿🇦 South Africa — 11 languagesBRICS-ready · Deterministic core · Local evidence
Brazil
Português
Russia
Русский
India
हिन्दी · ગુજરાતી
China
中文
South Africa
11 languages
FAQ · Trust

Frequently asked questions

Evidence-led answers with audit trails. No auto-approval.

No. It is a civic intelligence layer that clusters demand, fuses demographics and infrastructure gaps, and recommends candidate projects for human review. No auto-approval; every recommendation carries a human_review_notice.
Citizen PWA + Government dashboard · OPEN

Turn public voice into measurable action.

Evidence-led · Privacy-preserving · Human-governed. About JANSETU · Privacy · Terms

Deterministic v1 Centroids only
Try the demo — citizen + gov
Create accountLog in
No account to preview · Gujarati, Hindi, English

Explore: How It Works · Impact · BRICS · Contact · API · Sitemap

© 2026 JANSETU AI · Citizen-first · Evidence-first