Tenure paperwork for rural areas and informal settlements in South Africa
Your home is real. Now the paper can prove it.
Community facilitators photograph the affidavits, letters and structures a household already has, and pin the plot with GPS. Tenure Vault reads them, checks them, lists what is missing and drafts the letters. An official always makes the decision.
Household file readiness
Illustrative scores that move on a timer. Not live telemetry.
What this does not do: it does not issue title deeds, does not decide ownership, and does not replace the Department of Human Settlements or traditional councils. It prepares files. People decide.
The problem
Households without a formal deed or proof of tenure struggle to get finance, services and investment. Reported backlogs run past 1.2 million title deeds (see Research).
- Home is built
- No deed issued
- No proof for banks
- Dispute or fraud risk
- Eviction threat
- Claim stalls
- Documented and resolved
Complete file
Documents are readable, consistent and the plot is pinned. Ready for an official to review.
Gaps to close
Something is missing, blurry or inconsistent. The gap report says what to collect next.
Needs authority review
Signals such as a mismatched name or edited scan. A person investigates before anything moves.
The system produces leads and drafts for human review. It never produces certified determinations of ownership.
How it works
- Capture. A facilitator photographs documents, affidavits and the structure on a phone, offline. GPS photos and beacons or RFID tags mark the boundary.
- Read. OCR, a quality check and fraud-signal detection run on each image.
- Extract. Names, dates, plot descriptions and signatories become structured fields.
- Cross-check. Fields are compared with available open data.
- Gap report. A plain list of what is missing or inconsistent.
- Draft. Standard supporting affidavits and letters to the Department of Human Settlements or a traditional council.
- Human decides. An officer approves, modifies, rejects, asks for a field check or escalates.
AI and people
Each model does one job. Model names below are as specified in the project brief.
Image 2
OCR, image quality assessment and fraud-signal detection on captured documents.
GPT-6 Luna
Extracts structured data from the OCR text.
GPT-6 Sol
Cross-references open data and writes gap reports.
GPT-6 Astra
Drafts supporting affidavits and letters to departments or councils.
On-device OCR
NVIDIA-accelerated OCR so capture works offline in the field.
Human reviewer (mandatory)
A housing officer or council secretary makes every final decision. This step is never skipped.
Outputs are probabilistic and can be wrong. Confirm model identifiers and availability against current provider documentation (Azure, AWS, Google, OpenAI) before relying on them. AI outputs in this prototype are simulated to show workflow. No cost range is published yet; it will be set after pilot scoping.
3D model
Drag to rotate, scroll or pinch to zoom. Move the slider to change how complete the household file is.
- House: the home people live in.
- Paper document: proof of who owns it. The red stamp appears as the file completes.
- Phone: takes clear pictures of the paper.
- GPS pin: marks the exact place of the house.
- Digital box: keeps the proof safe online. It glows as it fills.
- Round figure: the facilitator, who collects, photographs, then delivers to the box.
Simplified for teaching. It is not a survey or legal drawing.
Research and evidence
Households without deeds face finance barriers, eviction risk and long claim delays. These sources were supplied with the project brief; open each to read the original.
Business Day. National backlog above 1.2 million deeds.
Pro Bono. Missing deeds bring eviction risk and fraud exposure.
SERI. How title delays harm women and later generations.
Parliament. Systemic delays in land and title processes.
News24. Long-unresolved claims that need better documentation.
Coverage
Target areas with current weather from Open-Meteo, fetched in your browser and refreshed every 5 minutes. Weather matters for field capture days. The stage labels are the company's own designations, not from the feed.
About DNVS
Tenure Vault is built by DNVS, a South African private company working to turn the papers households already hold into proof of home.
Registration
| Company | DNVS |
|---|---|
| Registration no. | 2026 / 779165 / 07 |
| Type / date / status | Private company / registered 30 September 2026 / In business |
| Address | 10 Saul Jacobs Street, Mindalore, Krugersdorp, Gauteng, 1739 |
| Financial year end | September |
Our teams
General
info@dnvs.co.zaDevelopers
developers@dnvs.co.zaAI solutions
aisolution@dnvs.co.zaFinance
finances@dnvs.co.zaAdmin
admin@dnvs.co.zaThuso Rankgoma Dammie
Director
I started DNVS because too many families live in a home they built, paid for and cared for, yet cannot prove it is theirs. That gap closes doors to finance, services and security. I want Tenure Vault to turn the papers people already have into proof they can stand on, so that no household is left invisible. If we do this well, a family can borrow, build and pass on a home with confidence. That is the impact I want to make, and it is why I come to work every day.