Working software, every Friday.

1. PheroChain

   Vendor verification for procurement and finance teams in India. It runs PAN, GST, MSME, CIN, sanctions, court-record and bank checks on a supplier before a purchase order is raised, then re-checks every vendor every 30 days.

   Live, access by request · pherochain.com

   

2. NeverCram

   A study app for students. Paste notes, upload a PDF or drop a lecture link, and it builds flashcards and study sessions scheduled by the FSRS-6 algorithm. Free plan; Pro is $9 a month.

   Live since March 2026 · nevercram.app

   

3. YeloChess

   Play chess with a friend from a shared link, no account needed. Daily, rush, streak, themed and rating-based puzzles, plus a board for analysing positions. Free.

   Live · yelochess.com

   

Three ways to work with us.

Every engagement starts with a 30-minute call. We tell you which of these fits. If none of them do, we say so.

You'll know what's happening every week. Usually every day.

1. Day 0

   A 30-minute call

   No slides. You describe the problem, we ask questions, and we tell you which of the three options fits. Sometimes the answer is none of them, and we say so.

2. Day 3

   A written scope and a price

   Two to four pages. What we'll build, what we won't, what it costs, and when you'll see it. You edit it until it's right.

3. Every Friday

   A demo of working software

   A URL you can open. We record it so anyone who missed it can watch. You tell us what to change and we change it the next week.

4. Last day

   Handover, then 30 days of fixes

   Repo, docs, credentials, and a short list of what we'd do next. Anything that breaks in the first 30 days, we fix at no charge.

What AI-native means here.

Four things we do on every project, in enough detail that you could audit them.

1. Agents write the first draft. People sign off on every line.

   Each of us works with coding agents all day. We write the spec, the agent writes the first pass, and we review the diff like we'd review a colleague's. That's where the speed comes from, and it's why a small studio can ship what used to take a team of ten.

2. An agent tries to break it before you see it.

   Before a Friday demo, an agent runs through the feature looking for the things we missed: empty states, bad inputs, slow queries. It finds some. We fix them. You see fewer bugs.

3. AI goes into your product only when it earns its place.

   Most features don't need a model. When one does, we write an eval first, test three or four models against it, and pick the cheapest one that passes. You see the eval and the cost per user before we commit.

4. You can audit all of it.

   We keep a written log of what the agents did on your project. If you want a part of the codebase done by hand only, say so and we'll do it that way.

Questions people ask first.

Where are you based, and what hours do you work?

We're in India, on IST (UTC+5:30). For US clients that means your morning is our evening. We keep at least four hours of overlap with your working day and we're on your Slack during it.

How do you price work?

Fixed price for sprints and builds, agreed in writing before we start. A flat monthly rate for embedded engineers. We don't bill by the hour. A slow day is our cost.

Who owns the code?

You do, from the first commit. We build in a repo under your account, or we transfer it on day one. There is no lock-in and no licence fee.

You use AI to write code. Is that safe for my IP?

We use commercial models with zero data retention agreements, and we can use your own API keys so the data stays under your account. Every line an agent writes gets reviewed by a person before it's merged. If you'd rather we didn't use AI on a specific part of your product, tell us and we won't.

What if I already have an engineering team?

Then the Embedded Team option is probably the right fit. Our engineers join your process rather than replacing it. Or start with an AI Sprint on one feature your team hasn't had time for.

What do you build with?

Mostly TypeScript: Next.js or React on the front end, Node or Python on the back end, Postgres, and whichever cloud you already pay for. For AI features we work with Anthropic, OpenAI and open-weight models, and pick the cheapest one that passes the eval.

Do you sign NDAs?

Yes. Send yours, or use ours. We sign before the first real conversation if you want.

How do payments work from outside India?

We invoice in USD or INR. Bank transfer or card. We're a registered private limited company, so you get proper invoices your accountant will accept.

Anything not here, email info@gigglestack.com. A person replies within one working day.