Community Wolf (@CommunityWolfAI) header
Community Wolf (@CommunityWolfAI) profile picture

Predicting crime. Powering safety. Building infrastructure.

Johannesburg, South Africa
Joined September 2024
One network. Four signal types. Most safety data has a single-source blind spot. SIA composes four into one.
1
11
There is a weather API. A payments API. A maps API. There has never been one for physical safety. So every team built it themselves or made do with stale public crime data. SIA changes that.
3
3
33
SIA is the Safety Intelligence API. One call returns a safety score for a location, the breakdown behind it, and the context around it. We own the data layer. You query it.
100
A mother in Joburg pulls up Uber, sees the surge, decides to drive her daughter herself. The route options come up. Twelve minutes either way.
1
38
Insurance is, at its core, a pricing problem. You charge for the risk you can see. The risk you can't see, you average, and someone always pays too much while someone else pays too little.
1
11
Every logistics route in SA is a risk decision. JHB CBD to Roodepoort at 3am ≠ 11am. Materially different, different hijacking incidence, different no-go zones, different escalation patterns. Most fleet platforms can't see any of that.
1
24
If you're building an AI agent and you want it to know whether a street in Joburg is safe right now: don't scrape news sites. Don't pay a data broker. Point it at SIA and call one tool: query.hex_score.
1
2
26
We just made it possible to build on physical safety data the same way you build on Stripe, Mapbox or Twilio. SIA, our Safety Intelligence API, is live in production.
1
2
44
A grandmother in a township sees a suspicious vehicle outside her grandson's school. She sends a WhatsApp voice note in isiXhosa. I n 90 seconds: transcribed, categorised, geo-tagged, alerts sent, the safety map updates.
1
2
28
1 security business. 6 months. 528,000+ licence disc scans. 269 flagged vehicles. Zero new hardware.
1
6
The most underrated thing about safety data is who's allowed to contribute to it. For decades the answer has been: nobody. SAPS keeps theirs. Insurers keep theirs. Every WhatsApp group keeps theirs.
1
2
13
As you drive through the city, as your GPS transfers from one hex to the next, it pings our API and tells you what the risk is. That's the vision.
🤖 Made with AI
1
2
20
Five layers of safety intelligence. Sampled every 500m. Layer 1: Precinct data (45%) Layer 2: Crime hex density (20%) Layer 3: Community reports (15%) Layer 4: Live threats (15%) Layer 5: POI proximity (5%)
1
14
From clipboards to crime intelligence. That's the journey this industry is on.
1
1
8
The same neighbourhood has a different safety score depending on what you're doing there. A suburb can be safe to live in but risky to drive through at night.
1
1
10
12,000 patrol events. Tracked, timestamped, GPS-verified. Here's what the data shows:
1
2
26
1,150,000 WhatsApp messages processed by AI. Not a projection. That's what's happened.
1
1
18
6,000 safety reports. 7,500 community users. 4,800 monitored areas. Every report goes on the map; geo-tagged, categorised, and scored.
1
1
10
The most dangerous route between two points isn't always through the worst neighbourhood. It's the one at the wrong time.
1
7