Crawler rules
robots.txt allow rules for OAI-SearchBot, ChatGPT-User, PerplexityBot, Claude-SearchBot and Applebot-Extended, with cart routes closed to scrapers.
stacktxt builds the layer agents read: crawler rules, plain-text manifests, structured data and product feeds, delivered to your site so AI agents can see, use and buy from your brand.
From crawlable to shoppable
Start with what crawlers read. Add your catalog next, then real-time sync and checkout as agents start to buy.
-
Stage 1 · Starter, $99/mo
A Lighthouse audit, robots.txt allow rules for AI search crawlers, a /llms.txt manifest and an ARD manifest at /.well-known/ard.json. No setup fee.
-
Stage 2 · Growth, $499/mo
Everything in Stage 1, plus a normalized product feed, /llms-full.txt, Product and Offer JSON-LD, a daily feed push to ChatGPT Business Hub and Google Merchant Center, and regional demand by DMA. No setup fee.
- <500ms[Catalog change to edge cache]
Stage 3 · Scale, $1,499/mo
Everything in Stages 1 and 2, plus edge webhooks that sync catalog changes to Cloudflare KV, multi-region manifests, UCP cart endpoints and weekly checkout simulations. One-time setup fee.
Every file an agent asks for
We write, test and deploy the machine layer as a ZIP bundle, through your Cloudflare zone or as a pull request, and push your feeds to the shopping platforms.
See how we deliver[New · Nov 2026] Business Hub feed push
A daily server-to-server push of your catalog to your ChatGPT Business Hub account as jsonl.gz, tsv.gz or Parquet, so your catalog stays inside OpenAI’s 14-day retention window.
Plain-text manifests
/llms.txt and /llms-full.txt strip markup and scripts, so a model reads your catalog with 90 to 95% fewer tokens.
WebMCP search tools
Search tools registered in document.modelContext, so an agent on your site can query inventory directly.
Details across
the three stages
See all deliverables -
ARD manifest
A discovery file at /.well-known/ard.json that points agents to everything else.
-
Structured data
Product, Offer, return and shipping JSON-LD with zero Rich Results errors.
-
Lighthouse audit
Page speed and Core Web Vitals, measured the way a crawler loads your pages.
-
UCP readiness
Whether an agent can reserve a cart through /.well-known/ucp.
-
Regional demand
Where your visits from chatgpt.com and perplexity.ai come from, by region.
-
Edge sync
Catalog changes pushed to Cloudflare KV in under 500ms.
“A client quote goes here once they approve it.”
[Delivery] How the files
How the files
ship
Your developer uploads a ZIP, you invite us to Cloudflare, or we open a pull request. Feeds go to ChatGPT Business Hub and Google Merchant Center.
ZIP bundle
The robots.txt patch, /llms.txt, /llms-full.txt, ard.json and feeds in one bundle for your web root.
Cloudflare Worker
Invite [email protected] to your zone and we serve the machine routes from the edge.
Pull request
For Vercel, Netlify or theme repos, we open a branch with the manifests ready to merge.
Daily feed push
Your catalog sent server to server to ChatGPT Business Hub and Google Merchant Center every day.
“A second client quote goes here.”
[How it starts]
From a first scan to monthly checks on what agents read
-
01 · Free scan
Enter a URL. We check robots.txt, /llms.txt, structured data and feeds the way an agent reads them.
-
02 · Audit report
What agents can and cannot read on your site, with the fix for each gap.
-
03 · Deploy
We ship the files as a ZIP, through Cloudflare or as a pull request, then check every route.
-
04 · Monitor
Monthly drift checks, crawler triage and a short email on what changed.
Never start from a blank robots.txt
Every stage ships as plain files you own. Read them, keep them, move them.