Built by Us
We don't just advise. We build and operate real AI systems in production. These are our proof points — the working systems that keep our expertise sharp and operating-grade.
shilpiworks
Human imagination. Machine precision.
shilpiworks is an AI-powered creative commerce platform where 16+ autonomous agents continuously generate, validate, price, and publish wisdom stickers, bookmarks, and laser-cut art. The entire pipeline — from theme selection through image generation, OCR validation, and smart pricing to auto-publish — runs on production cron schedules without human intervention.
Built on Next.js, Prisma, Vercel Postgres, and the Mastra agent framework, shilpiworks is a living laboratory for the same AI agent patterns we advise clients on: multi-model orchestration, decision-trace capture, outcome linking, and autonomous operations at scale.
The stickers themselves are “wisdom stickers for readers, leaders, and lifelong learners” — drawing from philosophy, science, literature, indigenous knowledge, and culture. Thirteen curated collections span themes from Stoic philosophy to curious kids to Minnesota wildlife.
What Makes It Work
The technical architecture behind shilpiworks reflects the same patterns we help clients design and deploy.
16+ Autonomous Agents
Themed AI agents (Stoic philosophy, Indigenous wisdom, Science, Women of Wisdom, and more) running on production cron schedules — zero human designers.
Multi-Model AI Pipeline
Gemini 2.0 Flash for speed and cost, GPT-Image-1 for generation, with intelligent fallback and cost optimization across models.
Decision Trace Capture
Every agent run logs its full reasoning chain — theme candidates, style selection, OCR attempts, quality scores. Institutional memory, not black boxes.
Outcome Linking
Agent decisions are connected to product performance — page views, cart adds, purchases, and revenue. The system learns what works.
OCR Validation Pipeline
Text legibility is verified before publishing. Failed stickers are auto-retried with corrections — no broken products reach customers.
Perceptual Hashing
Near-duplicate detection prevents the agent network from repeating itself. Theme diversity is tracked and enforced across all agents.
Engineering Dispatches
The engineering story behind shilpiworks, told through blog posts as we build.
Your Database Is Fine — The Proxy Isn't: A Prisma Data Platform Outage Story
The Prisma Data Platform proxy went unreachable during a routine deploy. The database was healthy the entire time. Here is what we learned about deprecated middlemen and why ISR caching saved the site.
Organic Engagement Loops for an AI Sticker Shop — What We Actually Built
27 AI agents, 600+ stickers, and zero engagement surface. The site architecture was one-directional — agents produced, visitors browsed, nothing flowed back. Here is how we added hearts, shares, a fresh feed, and trending — and why it matters for the context graph.
Curation at the Edge: The Strategic Value of Domain-Specific AI Constraints
The ECE agent isn't just generating art; it's automating the curation of a technical canon. By applying domain-specific constraints—circuit-schematic aesthetics and curated pioneer lists—we transform generic generative AI into an authentic practitioner's tool.
Engineering Wisdom as a System: From Blueprints to Binaries
Engineering wisdom is a high-entropy dataset. To make it consumable for the Shilpiworks Engineering Collection, we developed a 'Blueprint Technical' visual system that provides a low-entropy interface for complex technical insights.
Executive Verdict: The Chat-Based Control Plane is a Systems Trust Boundary
A systems-level diagnosis of using Telegram as a production control plane—balancing operator proximity with trust boundary rigor.
Why Generic AI Falls Short for Technical Audiences: Building the ECE Sticker Agent
Generic quote generators produce generic results. When we wanted stickers that actually resonated with electrical and computer engineers, we had to build a dedicated agent with domain-specific visual language—circuit-schematic aesthetics, monospace typography, and a curated canon of EE/CE pioneers from Claude Shannon to Jensen Huang.
We Wiped Our Production Database in 90 Minutes: A Prisma Postgres Post-Mortem
A routine analytics feature addition turned into a full production database wipe. Here's how running raw SQL through Prisma Postgres destroyed 590 products, how we recovered from a backup, and the rules we now follow to make sure it never happens again.
Two Silent Killers: Vercel Cron Auth and the Two Flavors of OpenAI 429
Five of our eight AI agents silently stopped running for days. The culprit was two separate bugs that looked identical in the logs — one a misconfigured auth header, the other a billing cliff. Here's how we untangled them.
Building a 712-Species Wildlife Sticker Machine: Dataset-Driven AI Generation
We took the Minnesota DNR's public dataset of 712 wildlife species and wired it directly into an AI agent. Every day, it picks the next ungenerated species, researches it, generates a dreamy watercolor sticker, and publishes it automatically. Here's what we learned about dataset-driven creative pipelines.
The Vercel Edge Cache Trap: Why force-dynamic Doesn't Always Work
We kept adding new stickers to the shop and they wouldn't appear — even after redeploying. The page was serving 8-day-old HTML. We tried everything. Here's the root cause and the one-line fix.
We Built an AI That Watches Our AI: The Ops Observer
We have 7 AI agents running 24/7 generating stickers. When they fail silently, we don't know until we check the database manually. So we built an eighth agent whose only job is to watch the other seven — and email us a diagnosis when something goes wrong.
When AI Generates Art: The Hidden Challenge of OCR and Transparency
We launched a new autonomous agent to generate Minnesota wildlife stickers every day. Here is what we learned when the AI decided to paint our text instead of writing it.
Debugging Mastra: Why Our AI Workflow Silently Ate Errors
We built six specialized AI sticker agents on Mastra workflows. They all silently stopped publishing. Here's the root cause, the debugging expedition, and three rules every Mastra user should know.
The same patterns, applied to your business
The AI agent architecture, experiment design, and operating integration we use in shilpiworks is what we help clients build. The lessons from operating feed directly into the advice we give.
See How We Can Help