I Lost My Favorite Software. So I Built My Own.
From 2008 - 2021 , I was running a product company in Sweden. We used TradeGecko for everything, inventory, purchase orders, sales orders, invoices. It was perfect for a small team.
Then Intuit acquired it in 2020, rebranded it to "QuickBooks Commerce," and killed it completely in 2022. Over 10,000 businesses were left scrambling.
I tried everything after that. Cin7, powerful but built for enterprise. Zoho Inventory, decent but buried inside an ecosystem of 50 apps. Spreadsheets, we all know how that ends.
Nothing felt like TradeGecko. Simple, clean, built for small product businesses. So in March 2025, I started building VNDLY.
What it does:
• Inventory, sales orders, purchase orders, invoices — all in one place
• Product Development Pipeline, Packing Visualization, Sales & Leads Strategy Planning, a built in Year Wheel and team Chat.
• AI assistant that generates reports, forecasts demand, and spots anomalies
• Multi-location stock tracking with real-time levels
• Multi-currency Price lists, packing lists, PDF exports
• Integrations: Shopify, WooCommerce, BigCommerce, Squarespace, XERO, Quickbooks, Visma, Fortnox, Shippo, EasyPost, Sendcloud, etc.
• Starts at $49/mo with a 14-day free trial
The twist: I'm not a seasoned developer. I built the entire thing using my limited knowledge and AI tools like Claude and openAI. I'm a designer and business guy who learned to ship software with AI as my co-pilot and I have learnt a lot the past year for sure.
I sold my previous company in 2021 and after some Real Estate business ventures I am now I'm going solo on this. No funding, no team, just me and a laptop in Stockholm.
Where I'm at:
• Live at vndly.io
• Just launched on Product Hunt
• Podcast running (VNDLY Insights on Spotify)
• Blog, posting articles every weekday.
• Currently working on marketing strategies, google ads, reddit ads, etc.
I'd love feedback from anyone who's dealt with inventory management pain. And if you're a TradeGecko refugee, I built this for you.
Great story! TradeGecko left a real gap for small product businesses.
Since VNDLY sits in the operational core of a company, security and data integrity will matter a lot for adoption.
A few areas users will likely ask about:
• If you connect Shopify, WooCommerce, or accounting systems, how are API tokens stored and encrypted?
• With the AI assistant generating reports and forecasts, are financial or inventory datasets sent to external AI providers?
Inventory platforms hold sensitive business information. Sales volumes, suppliers, margins, and stock levels. Clear boundaries around data handling will build trust with small teams considering migration.
Another useful signal is transparency:
• What data is stored
• How integrations authenticate
• How backups and recovery work
If you position VNDLY as the operational hub for product businesses, a simple security and architecture page will help technical buyers feel comfortable moving their inventory and financial data.
As a security team building Nautillo Pro, we often test platforms with many integrations because API connections expand the attack surface quickly.
This is incredibly valuable feedback, thank you! You hit the nail on the head, when a platform sits at the operational core of a business, trust is the absolute highest currency.
Because I was building this mostly solo, I knew I couldn't afford a single breach or data leak. Here is exactly how we handle the security and data boundaries you mentioned:
1. Integration API Tokens (Shopify, Woo, Xero, etc.)
API tokens are never stored in plain text. We use a dedicated encryption utility (
AES-256-GCM) to encrypt all access tokens before they ever touch the database. The encryption utilizes a secure initialization vector (iv) and an auth tag for validation, and the master decryption key (AI_KEY_SECRET) only exists as a secured environment variable on the server itself. Even if someone were to gain read access to the database, the tokens in theintegrationstable are completely unusable without the server-side memory key.2. AI Data Boundaries & External Providers
This was a huge concern of mine, which is why we went with a BYOK (Bring Your Own Key) architecture for the AI.
VNDLY does not have a central, monolithic OpenAI account that ingests all tenant data. Instead, the user provides their own API key (which is also encrypted via AES-256 in the database). When they ask the AI a question about their inventory or sales, the request is executed using their key, directly to the provider. We don't train our own models on their data, and because they use their own key, they retain the data privacy agreements provided by OpenAI/Anthropic for API users (where data is not used for model training).
3. Tenant Data Isolation
On the database side, we use PostgreSQL Row Level Security (RLS) policies. Every single query automatically enforces a strict tenant boundary (
tenant_id = get_current_tenant_id()). It is mathematically impossible for a query executed by one company's user to retrieve records belonging to another company.You are completely right about transparency. Adding a dedicated "Security & Architecture" page to our website that breaks this down in plain English is a brilliant idea, and I am going to build that out this week.
It's great to hear from a security pro! If you have any other thoughts on how small SaaS platforms can build trust, I'm all ears.