Three tools. Real problems, solved.
Invoice Consolidation Tool
Importers filing IDFs on KRA's iCMS system often deal with invoices containing hundreds of line items — many of them near-duplicates (same product, different color or minor variation). This increases the risk of mismatches, rejections, and slow manual review.
This tool takes a commercial invoice (PDF, Excel, or CSV) and automatically consolidates it down to under 99 line items, as required for iCMS filing. It groups items by HS code, description, origin, and unit of measure — so items like "Hose Pipe Blue" and "Hose Pipe Yellow" become "Hose Pipes ×2," while keeping genuinely distinct items separate.
Every consolidation is checked against the original invoice to guarantee the totals match exactly — nothing is lost or altered in value, only grouped. The result is a clean, compliant, audit-ready invoice, ready for IDF filing, cutting manual prep time and reducing rejection risk.
Frequently asked questions
Product screenshots
Sheetsync
Many business owners track M-Pesa spending manually — going through SMS messages one by one to log expenses. Sheetsync automates this.
The app reads incoming M-Pesa messages, automatically extracts the amount, recipient, transaction type, and time, and organizes them into categories (fuel, supplies, utilities, etc.) based on recipient name keywords. Users can manually override the category on any transaction, and overrides sync back immediately.
Every transaction syncs automatically to the user's own Smartsheet — no manual data entry. The result is a live, organized record of business spending that updates itself.
Frequently asked questions
Product screenshots
selfie.ke — Loyalty Points Automation
selfie.ke ran two offline systems at the till: Microsoft Dynamics RMS 2005 for payment and receipts, and Hero Points v3.6.1.0, a legacy add-on with no API and no way to talk to anything external — not WhatsApp, not a spreadsheet, not a phone. The client wasn't willing to touch Dynamics itself, so loyalty tracking was stuck: no automated reward notifications, no audit trail, and roughly 9,500 existing customers' point balances sitting in a system with no export path anyone had actually tested.
We built an internal Django app used by till staff alongside Dynamics — never integrated with it. Staff search for a customer, enter the sale amount, and points are calculated and added automatically, with no field anywhere letting a person type a points value directly. Crossing a reward tier triggers an on-screen celebration and queues a same-evening WhatsApp notification via Meta's Business Cloud API, and marking a gift collected fires a second confirmation automatically.
Points are structurally un-editable — a Postgres trigger rejects any write outside the two functions that calculate them, and a second trigger blocks reversing a collected reward. Migration was validate-then-import, never a blind load: a dry-run report caught dirty data before a single row was written, and historic balances were deliberately not retroactively rewarded, the decision that avoided roughly 1,626 gifts being owed on day one. The result: ~9,500 customers migrated with zero incorrect payouts, and a full audit trail where none existed before.