The website finally shows the work: four aisles and an allowlist
We rebuilt decisionsciencecorp.com around what we already ship: four engagement aisles, blog proof on an allowlist, and a buy line that only has to be said once.
Read more →Blog
Applied data science insights, signal extraction techniques, and analysis methods.
Pro Spike Flow, June: About Nineteen Times the Bill
In June 2026, Pro Spike Flow's store ran about nineteen times what we billed them for the month. Hit Harder is the platform we built to make numbers like that reproducible across…
Read more →Why we shipped search-users before Phase 3 finished
Ask Q Phase 3 ships in pieces. The piece that shipped today is search-users for the chatter lane — the boring fix that ends a class of agent failure: pasting UUIDs into a Tasks…
Read more →Invoicing: Parsedown, SITE_URL, and the boring 2%
Money software loses trust on the parts most teams never test — markdown rendering and link resolution. We fixed both: Parsedown renders the breakdown, SITE_URL points share…
Read more →Square round-trip: in-person sales on the same order board
The one-menu-database post described the happy path: website, Square catalog, and DoorDash reading the same menu rows in Kitchen POS. That story was true and incomplete.
Read more →Decision Science Corp runs on Sanctum Tasks: what our own client adoption looks like
SanctumOS told the platform story; this is the client half. Decision Science Corp runs daily operations on Sanctum Tasks—not a demo tenant, the shop that builds Sanctum.
Read more →Channel partners on a ghost kitchen: affiliate pickup without a second menu
Delivery apps want you to believe the only way to sell food online is through their tile on a phone screen. That is true for a lot of volume — and at Empanada Empire we still…
Read more →Hardening Ask Q: making a live Vernal webchat integration survive all-day use
Ask Q was already working. That is the part people forget when they read a hardening post and imagine a launch story.
Read more →Ask Porter: a partner-scoped Vernal co-pilot on Kitchen POS
Channel partners got a portal, a storefront, and recurring pickup windows in June. The next question affiliates ask is not philosophical — it is operational: *"How do I open my wi
Read more →Making CRM look like it came from the same shop as Sanctum Tasks
Sanctum Tasks had a design system. CRM had a generation gap. For a while, that was the state of things, and we lived with it - the way you live with a sibling who showed up to…
Read more →Kitchen POS operator UI: when the line caught up to Sanctum Tasks
In early June we published Making CRM look like it came from the same shop as Sanctum Tasks — the story of visual parity across CRM and Sanctum Tasks. Kitche
Read more →QR Code Studio: open-source QR generation with an API agents can actually use
It started with flyers and App Store QR generators full of ads. We built qr.decisionsciencecorp.com for family — then open-sourced the API, agents.md, and SMCP plugin.
Read more →From Ten Years to Three Days: Why Our Mailbox Pipeline Didn't Need More AI
We indexed nearly a million Gmail messages and almost sent 35,000 senders through Venice enrichment at ten per day - a decade-long backlog. Most of the win came from email-native…
Read more →Sanctum Tasks June UI governance updates: skin prefs, org default, and admin lock
Sanctum Tasks users wanted to personalize the way their workspace looked. Skins, density, card spacing — fair asks, and ones we agreed with. The risk was the other side of the…
Read more →Warehouse scan, one month in: what the floor told us to fix
A month after barcode scan intake went live at Luke's, real floor use taught us four specific things the spec did not. The June 4 fix closed each one - and the lesson is the same…
Read more →CRM contact ingestion and enrichment controls: what shipped in May
Our CRM was drifting from reality. Phone backups held the real contacts — every number, every name, every recent connection — but they lived in iCloud and Google exports, never…
Read more →Inside the DSC blog CMS hardening sprint in May
Our blog CMS was shipping on schedule, but the seams were showing. Posts went up fine. What was rough was the operational surface around publishing: a bad post took a database…
Read more →Warehouse scan shipped at Luke's: schema, mobile flow, and coverage
Luke's warehouse intake had outgrown manual handling. The process worked when volume was low, but scan events, container moves, and assignment updates were too easy to lose in…
Read more →Document library and directory paths in Sanctum Tasks
After we shipped the core task API and then added comments and attachments in v1, we realized something was still missing: a way to organize long-form content. Tasks could have…
Read more →CRM refresh: what we shipped in May
Nineteen changes shipped in May, all driven by the same complaint from our team: I cannot answer that question without exporting to a spreadsheet. Here is what we did about it.
Read more →Restaurant Meta paid social: creative testing frameworks (teaser)
Meta paid social punishes restaurants that test one ad at a time. Here are the creative-testing frameworks that keep your CPM from doubling.
Read more →DoorDash cold start for a brand with zero reviews
Every DoorDash brand starts at zero reviews. The first sixty days are the most fragile period on the platform. Here is what works during cold start and what burns budget.
Read more →Shopify PDP research: YouTube corpus to ProSpikeFlow dev subdomain
56 practitioner YouTube videos → consensus-tagged PDP spec → sticky add-to-cart on prospikeflow.client.decisionsciencecorp.com. Methodology and phase-1 implementation in one post.
Read more →Five DoorDash takeaways for ghost kitchens
We've been synthesizing the same kind of analysis we did on Google Ads for restaurant marketing, this time for DoorDash. Ghost kitchens — delivery-only restaurant operations with…
Read more →Why we run client apps on a multihost pattern
One bad deploy taking down two production apps is a failure of architecture, not a developer. Here is the multihost pattern we run, and why it keeps every blast radius small.
Read more →Off-platform paid ads plan for restaurants (executive summary)
A restaurant that depends entirely on DoorDash and Uber Eats is one algorithm change away from a bad quarter. Off-platform paid ads are how you build demand the marketplaces…
Read more →Bulk Tasks API: creating 100 tasks without breaking rate limits
Sanctum Tasks was built for agents, not humans. That design principle shows most clearly when you need to create tasks in bulk. An agent creating one task at a time is slow…
Read more →Rate limiting and stable error codes in our Tasks API
When we built Sanctum Tasks, we built it for one specific audience: programs, not people. Every design decision — from the API structure to the error handling — was made with the…
Read more →Sanctum Tasks v1: tasks, comments, and attachments
Six days after we shipped the bare task API, we pushed v1 of Sanctum Tasks with the features that made it actually useful for real workflows: comments and attachments. This post…
Read more →Invoicing went live on multihost with Let's Encrypt in one afternoon
We launched our invoicing application with HTTPS and a functional deployment in one afternoon. That's a notable improvement from a few years ago when launching new services would…
Read more →Kitchen POS multi-location PRD: why v1 is deliberately narrow
v1 of Kitchen POS is deliberately single-location. Not because multi-location is hard, but because every v1 customer should be one we can support without rewriting the schema.
Read more →Invoicing app bootstrap in one day: repo, DNS, multihost, certbot
Every new web app needs four things before anyone can use it: a repo, DNS, a host, and TLS. Most teams treat that as a week. We treat it as one afternoon.
Read more →Native Letta on a temp VPS: no Docker, systemd, uv
We needed Letta running on a temporary VPS for a client demonstration environment. The hosting environment was minimal — basic Linux VPS, no containers, no systemd service…
Read more →Ghost kitchen marketing synthesis: 31 transcripts → one playbook
Thirty-one transcripts of ghost-kitchen marketing — founder talks, platform tutorials, interviews — synthesized into one playbook. We did the work for ourselves first.
Read more →Tool-message JSON pitfalls in Letta imports (and how we fixed them)
A routine Letta migration crashed our server. The .af file imported cleanly, then the API started 500-ing. The cause: tool-message JSON stored as plain English.
Read more →Migrating Letta agents without cloning Postgres: the .af workflow
Every Letta deployment eventually needs migration. You might be moving to a new server, cloning to a staging environment, consolidating multiple Letta instances, or simply…
Read more →Recovering 20 years from a ZIP: archive catalog at scale
We inherited a problem last quarter. A client handed us a 40-gigabyte ZIP file containing twenty years of their organization's digital history. Project files, documents, images…
Read more →Reading untrusted GitHub repos without running them
Job-seeker workflows have a steady hostile artefact in them: the "try our stack" GitHub repo. Some are bait. We have read several without running them, in a short-lived container…
Read more →How we wired Google Ads conversion tags before GA4 existed on a site
Running Google Ads without conversion tracking is like flying blind. You pay for clicks, but you never see what happens after the click. Did anyone actually order? Did they call…
Read more →What 50 hours of restaurant Google Ads YouTube taught us (summary)
Fifty hours of restaurant Google Ads YouTube, transcribed and synthesized into a single playbook. The summary you can act on before you spend real budget.
Read more →Convertible covers: from paper catalogs to vision-backed PDPs
They looked up every inventory item by hand in OEM catalogs. We built computer vision and catalog ingestion so pricing and fitment became programmatic—not manual.
Read more →ads.decisionsciencecorp.com: internal ads ops hub
Managing advertising campaigns for clients was becoming a mess. We had Google Ads in one dashboard, Meta in another, creative assets scattered across drives, and no single place…
Read more →"Bolt-on content" products: when WordPress is the right backbone
When simpler WordPress patterns beat a custom CMS—and when to stop. Fuel is our real bolt-on editorial example.
Read more →Migrating Fuel off DreamHost to LEMP
Fuel is a Roger Wilco product — the editorial people-management platform co-developed years ago by Mick Jones and Mark Hopkins — and it has run on DreamHost shared hosting since…
Read more →Reverse-engineering a legacy WordPress platform
Every WordPress site tells a story if you know how to read it. We recently took on a consulting engagement to reverse-engineer a legacy WordPress platform that had grown…
Read more →Ollama vs bare harness vs full agent stack (benchmark teaser)
Every team building with LLMs hits the same crossroads: how do you run inference? The options seem simple — run Ollama locally, use direct API calls, or wrap everything in an…
Read more →Phase-2 engine research: gemma spot-checks and routing notes
Every inference pipeline eventually needs a decision: do you stay with the current model or escalate to a larger one? We've been researching that question for months using…
Read more →Bumblebee smoke interview: latency, hallucination, format failures
Simple systems fail differently than complex ones. After months running Bumblebee in production: latency lies, matching hallucinates, format checks paper over bugs.
Read more →Running an agent harness on a $5 VPS
Can a $5-class VPS and a tiny local model do real agent work? CPU-only, no GPU. We ran the Bumblebee harness for weeks — here is what broke and what held.
Read more →Technonomicon → multi-agent weblog (plan teaser)
We're building something fundamentally different for our next phase of content operations: a multi-agent weblog where Idalia serves as editorial coordinator, managing specialized…
Read more →CRM.decisionsciencecorp.com: our take on a small-business CRM
Every small business needs a CRM. Nobody wants Salesforce's price for one. Here is the CRM we wanted — pipeline, leads, deals, reports, no enterprise pricing.
Read more →Kitchen POS: one menu database for web, Square, and delivery
Website, Square terminal, and DoorDash menus drift apart the moment three people own them. We gave all three the same source of truth. Here is how Kitchen POS does it.
Read more →Luke's Inventory: when spreadsheets break
Every small business has a spreadsheet moment. That's the moment when your spreadsheet stops being a tool and becomes a liability. We hit that moment with Luke, and it's the…
Read more →Payroll tooling: why we extracted payroll from spreadsheets
Payroll for a ten-person shop should not take a Friday afternoon. We threw out the spreadsheet, built a tool, open-sourced it. Here is what we learned.
Read more →Empanada Empire: our ghost-kitchen field experiment (why a data science company did this)
We opened a real ghost-kitchen brand in Richardson, TX — not as a pivot, but as a live lab for marketplaces, first-party ordering, and paid acquisition. Here is why the blog…
Read more →First-party ordering vs marketplace apps: architecture choices
DoorDash takes thirty percent. Your own website takes payment processing. The first-party vs marketplace decision is a margin decision — and most restaurants get it wrong.
Read more →Orderboard: the ghost-kitchen pickup display we open-sourced
Orderboard is not project-management software. It is a wall-mounted, driver-facing pickup board for delivery-only kitchens — with a REST API, Python SDK, and an agent plugin so…
Read more →Project management agents can actually use
Every PM tool we tried treated the API as an afterthought. So we built Sanctum Tasks — project management for programs, not boards or draggable cards.
Read more →Inventory ops for a warehouse auction
A warehouse-scale data-science engagement: organize auto parts and upholstery, vision pipelines and Sanctum agents, and deploy listings to auction sites rapidly.
Read more →