Each product we make is an experiment to understand customer's needs. We learn, we iterate, we ship.
Gherkin-driven test traceability across unit, integration, and E2E layers. Know which business scenarios are actually tested.
Line coverage says 87%. Nobody can tell you if checkout is tested. SpecTracer bridges the gap between your Gherkin .feature files and the JUnit XML / Cucumber JSON your test suites already emit.
Point it at your feature files and test results. Get one self-contained HTML report showing which business scenarios are covered at which layer — with a schema-validated JSON twin for CI. No test runner, no agent, no database. One CLI, one config file, one HTML file out.
It works with pytest, behave, JUnit, TestNG, Jest, Playwright, Cypress, Cucumber, and dozens more. Make it a CI gate that fails the build when the pyramid inverts or coverage drops below your floor.
Open-source LLM tool-calling research harness. Design controlled experiments to understand how and why LLMs choose the tools they do.
LLMs make hundreds of tool-calling decisions. Llaboratory lets you design fake tools with static or dynamic responses, wire in any OpenAI-compatible model, and run one session or a thousand to see exactly what the model picks and why.
Assemble versioned, reproducible testing plans that pin tool versions and freeze model snapshots. Fire N repetitions in one click with Batch Runs and get a real sample size instead of chasing single-session anecdotes.
Per-session metrics, cross-model comparison, tool-selection rates, call-order patterns, CSV export, and downloadable markdown findings reports — all self-hostable via Docker Compose.
A client for nostr, a rapidly growing decentralized alternative to corporate-run social media.
This started as a casual and personal weekend-project by one of our founders. However, this became the first product of ours with wide adoption. Zapddit is open-sourced under MIT license.
While other nostr clients mimicked Twitter and allowed users to follow people, zapddit takes a fresh perspective and allowed users to follow topics (hashtags and communities) instead. Instead of just having a like button, zapddit provides allowed both upvoting and downvoting possible.
At Nostrasia 2023, Zapddit won the award for the "Empowering Communities" category.
A simple Shopify app that helps merchants boost their conversion rates.
Greeniemart is a simple Shopify app that helps merchants boost their conversion rates.
With this being one of our very recent products, we got a few installations and feedback too. With paid customers, we see this as a good prospect for our product-line.
This experience even motivated us to work on another Shopify app.
Your customers can specify who they're ordering for. Perfect for healthcare, gifts, or regulated items where compliance matters.
With "ProxyBuy - Buy on Behalf," your customers can specify who they're ordering for.
This info appears in your Shopify orders screen, making it easier to verify recipients when needed. It's perfect for healthcare, gifts, or regulated items where compliance matters.
Simple for your customers, helpful for you, and ensures smooth, compliant order management.
Face recognition as a service targeting banks, schools and hotels.
FaceQuest was created in 2020 as a 'Face recognition as a service' targeting banks, schools and hotels.
The MVP was created and the pandemic hit when it was ready to hit the market. Once lockdowns subsided, we couldn't get enough traction from businesses who wanted to work with us.
At a later stage, we discovered that FaceQuest was not a product that today's businesses absolutely needed. We were literally selling vitamins instead of pain-killers 😂
AI safety net for SaaS Sales teams to reduce human errors in the sales process.
Notesally was created in 2022 as an AI safety net for SaaS Sales teams. The goal was to reduce human errors in the sales process by automating the data entry into CRMs by reading entries from emails or Linkedin chats.
We had a decent launch on Product Hunt. We got a #10 day-rank, 100+ upvotes and great reception on Product Hunt.
Despite all of these positive signs, we struggled to get paid customers. Trials were many in number but this was not a product the customers could pay for. Also, the rise of great competitors in this space like GTMBuddy made it tough for us to continue.
With no revenue at the end of more than a year, we decided to wrap this product up.