SafeRide - Safer journeys for every woman

Safer journeys for every woman.

For survivors

Start a private report before the signal returns.

The mobile app creates a local draft first, then guides evidence, location, legal framing, and consent decisions without forcing an upload.

How it works

For communities

Four pathways: keep private, map anonymously, seek referral, or escalate.

SafeRide separates personal evidence from route-level signals so a rider can contribute to safety patterns without exposing names, audio, or exact journeys.

See the index

Open source

Auditable safety tools, with a first-party API roadmap.

The repo includes the Android client, legal content packs, local Gemma assistant scaffolding, and Supabase prototype docs while the backend moves toward a dedicated API.

View the stack

Safer rides for every woman.Your evidence. Your control.Report. Refer. Stay safe.Justice starts with one tap.

An open-source safety app for women in Nairobi's public transport.

Impact

Safety before scale.

These are product guardrails, not vanity metrics: SafeRide should first prove that a rider can keep control of what is captured, redacted, shared, and deleted.

forced uploads before a rider chooses a pathway

survivor-controlled pathways: private, map, referral, escalation

redaction modes before an escalation packet is shared

Capture evidence without signal, pressure, or shame

For survivors

Offline tips, legal tags, and referral choices

How it works

Route signals with survivor identity removed

Route Safety Index

Be a force for safer transport.

SafeRide route safety team inspecting a matatu before a pilot

PILOT READINESS

The app flow is already mapped from stealth entry to consent gate

Onboarding, evidence intake, What Happened, Where/When, Legal Framing, Pathway Selection, Consent Gate, and Case Tracker are implemented as mobile screens.

Partner with us

SafeRide Android dashboard with emergency, report, case tracker, chat, and learning actions

PRIVACY-FIRST AI

On-device Gemma support is scoped for guidance, transcription, and offence tagging

The local assistant registry documents a 4-bit Gemma-3n model with text generation, audio transcription, offence tagging, checksums, and a survivor-centred system prompt.

Read the model plan

SafeRide practical guidance and legal aid illustration

ACCOUNTABILITY DATA

Legal and referral content is local before it becomes a service dependency

The repo ships 20 tips, 15 legal tags, and 5 seed provider records covering medical care, P3 forms, 1195, GBV centres, legal aid, and evidence safety.

Explore the index

Community Spotlight

Signals reviewers can inspect.

1 / 4

SafeRide consent-first reporting workflow on an Android phone in a Nairobi transport context

Private reporting flow

Mobile App

"SafeRide creates a draft first, tracks completed reporting steps, and only moves to sharing after a pathway and consent review."

Fund the Pilot

Together we can make every matatu ride safe.

Join our global movement. Fund the SafeRide corridor pilot and help us publish the Route Safety Index for Nairobi.

Fund the Pilot Partner with Us