work/closeby/

CloseBy

Local shop platform for Guna, with AI tools that take routine work off shop owners

Solo: design, frontend, backend, AI layer · 2026

Next.js 15TypeScriptSupabase PostgresClerkClaude APIRazorpay

CloseBy: Local shop platform for Guna, with AI tools that take routine work off shop owners

Manual work before

Shop owners type every product by hand, keep stock in their heads, and answer the same buyer questions all day. Unpaid orders and refunds are chased by hand.

What it automates

-Stock entry from photos and voice lists
-Restock and price ideas
-Draft chat replies
-Cancelling unpaid orders
-Refunds for rejected orders

The workflow

triggerThe owner uploads a shelf photo or a voice list of stock.
step 1The AI reads the catalogue through tools and drafts a stock import.
step 2The owner reviews the draft and approves it. Nothing is saved before that.
step 3Buyers nearby find the shop through geohash search and place an order.
step 4The order moves through a seven-state machine. Razorpay holds stock until the server confirms payment.
step 5Unpaid orders cancel after 15 minutes. Rejected paid orders are refunded automatically.
outcomeCatalogue entry, order tracking, payment checks and refunds run on their own; the owner only approves.

status Live demo.

In short

-Every AI call passes one gateway: auth, rate limit, kill switch, daily spend cap, input cleaning, tool loop, cost logging.
-AI only drafts (stock imports, restock ideas, chat replies); the owner approves before anything is saved.
-Razorpay flow holds stock until payment is confirmed and auto-refunds rejected orders.
-27 eval cases covering Hindi aliases, misspellings and gibberish input.

Architecture

A Next.js app with a single server-side AI gateway. The model never writes to the database directly; it returns Zod-validated drafts that the owner approves.

System designOne Next.js app; every model call goes through the AI gateway, every order through the state machine.

user
client
guard
service
ai
data
external
call
logical link
← scroll →

01Owner uploads a shelf photo or voice list
02Gateway: auth → rate limit → kill switch → spend cap → input cleaning
03Claude tool loop returns a structured draft (validated with Zod)
04Owner reviews and approves the draft
05Approved items saved to Supabase Postgres; cost logged per call

Decisions

+ One gateway for every model call- Calling the model from each feature directly  Rate limits, spend caps and logging live in one place, so a new AI feature can't skip them.

+ Human approval before any write- Letting the model save stock automatically  A wrong price or quantity costs a real shop money; a review step is cheap.

+ Integer paise and a seven-state order machine- Float rupees and free-form status strings  No rounding errors, and illegal status jumps are rejected at the boundary.

Failure handling

failurehandling

Model returns malformed or partial output

Zod validation fails closed; the owner sees an error, nothing is saved.

Daily AI spend exceeds the cap

Gateway refuses new calls until the next day; manual entry still works.

Prompt injection in uploaded text

Input cleaning plus tool-only output; the model can't trigger writes.

fig.1 System designOne Next.js app; every model call goes through the AI gateway, every order through the state machine.

user
client
guard
service
ai
data
external
call
logical link
← scroll →

fig.2 System architectureEach layer only talks to the one below it. The model never touches the database.

users
Buyer
search nearby shops, order, pay, track
user
Shop owner
catalogue, orders, AI drafts to approve
user
Next.js 15 app
App Router · server actions · Clerk auth · role-gated dashboards
client
server
AI gateway
every model call goes through here
guard
Order service
7-state machine · integer paise
service
Realtime
live order updates · in-order chat
service
providers + data
Claude / OpenAI
tool loop · provider switchable
ai
Supabase Postgres
catalogue · orders · cost logs · RLS
data
Razorpay
checkout · webhooks · refunds
external

fig.3 AI draft pipelineWhat happens when an owner imports stock from a shelf photo or voice list.

Owner uploads a shelf photo or voice list
user
Session + role check
guard
not an owner → rejected
Rate limit → per-helper kill switch → daily spend cap
guard
cap reached → AI off until tomorrow, manual entry still works
Input cleaning + size limits
user text fenced against prompt injection
guard
Tool loop
model reads the catalogue through tools; can't write
ai
Zod validation
guard
malformed output → error shown, nothing saved
Draft shown to the owner
edit, approve or discard
client
on approve
Approved items saved
data
Cost logged for the run
data

fig.4 Order state machineTransitions are enforced in server actions; illegal jumps are rejected.

● PLACED
ACCEPTED
PREPARING
READY
COMPLETED

CANCELLED
unpaid orders cancelled after 15 minutes; stock released
REJECTED
owner rejects; paid orders refunded automatically

fig.5 Decisionswhat I rejected → what I chose, and why

Calling the model from each feature directly
One gateway for every model call
Rate limits, spend caps and logging live in one place, so a new AI feature can't skip them.
Letting the model save stock automatically
Human approval before any write
A wrong price or quantity costs a real shop money; a review step is cheap.
Float rupees and free-form status strings
Integer paise and a seven-state order machine
No rounding errors, and illegal status jumps are rejected at the boundary.

fig.6 Failure pathswhat can break → what the system does

Model returns malformed or partial output
Zod validation fails closed; the owner sees an error, nothing is saved.
Daily AI spend exceeds the cap
Gateway refuses new calls until the next day; manual entry still works.
Prompt injection in uploaded text
Input cleaning plus tool-only output; the model can't trigger writes.

Numbers

Eval cases

Order states

TODO

p50 import latency

measure and fill in

TODO

Cost per import

from cost logs

○ Grow the eval set to ~100 hard cases
○ Record a 90-second demo video
○ Stock substitution suggestions