Working Notes on Agent Systems/Brad Zhang

@teach_fireworks / X longform

Don’t treat impressions as traffic: I looked at getdesign.md and found that it is more like a programmatic SEO site that is very stable and worth learning.

Don’t treat impressions as traffic: I looked at getdesign.md and found that it is more like a programmatic SEO site that is very stable and worth learning. I first...

May 4, 2026 · 5 min read

Don’t treat impressions as traffic: I looked at getdesign.md and found that it is more like a programmatic SEO site that is very stable and worth learning.
Figure 1 / source image

Don’t treat impressions as traffic: I looked at getdesign.md and found that it is more like a programmatic SEO site that is very stable and worth learning. I first noticed this open source library with 70,000 stars for providing high-quality DESIGN.md. It follows the recent Google Stitch specifications: Then I found its official website. The overall style is more mature and it also has paid services. The overall link is probably: from the popular open source project to your own web site, providing paid options. This way is still good. I just finished reading it. So if someone opens his mouth and says, "How many thousands of traffic does this site have every month and how much does it earn every month?", you'd better put a question mark first. What I can confirm is that it does have investment credentials on its public page, and you can see a sentence of 1M+ monthly impressions in the Google search summary. But what is written here is impressions, not visits, and not paid conversions. So I looked it up and you can see the picture below. The current organic traffic of getdesign.md is not large, but the trend is growing; It mainly relies on traffic from India. The homepage and several brand landing pages are the core traffic entrances. The SEO strategy obviously favors the "brand word/competitive product word landing page" route. The picture shows the Top keywords in the India market: - get design: ranking 2, search volume 4.7K - lovable dev: ranking 16, search volume 13K - runway ml: ranking 28, search volume 103K - wise money transfer: ranking 20, search volume 3.9K - linear app: ranking 13, search volume 700 I looked at the homepage, list page, details page, about page and request Page, several key items are there: Home page, list page, and detail page all have clear title, description, canonical 1 Open Graph, Twitter Card 2 The page has WebSite and Organization structured data. 3 The details page has BreadcrumbList and CreativeWork. 4 robots.txt and sitemap.xml are both configured. 5 There are about 510 URLs in the sitemap. 6 The page HTML is not a pure front-end shell, and search engines can directly read the main content. This is a standard SEO website building idea and is worth learning. But the problem is also obvious. The biggest problem with this site is not that SEO has not been done, but that many pages are still stuck at "can be included" and have not steadily reached "can be captured on a large scale". There are three main reasons. First, the content of the details page is thin. For example, on a certain brand page, the title is Design System Inspired by Vercel, and the description is basically a short style summary. It can certainly be indexed, but the incremental information is not that large. For search engines, this type of page can easily fall into an awkward position: it can be included, but it may not be willing to give a high ranking. Second, it has a bit of a "double route" tendency. I can see from the sitemap at the same time: /design-md/vercel /vercel/design-md This is not necessarily a serious accident, but it shows that this site has swayed in the URL design. As long as the processing of canonical, internal links and sitemaps is not strong enough, it will be easy for weight dispersion, waste of crawling budget, or the same topic pages to compete for words. Third, its business page is not ruthless enough. I can confirm on the request page that it has a clear fee: $39 $49, and the copywriting is very clear: no platform, no monthly fees. In other words, it is now more like a one-time paid light product, or a semi-product service entrance, not like a subscription SaaS. It’s not that you can’t make money this way, but you have to accept the reality: if the traffic does not continue to increase, the income ceiling will not be particularly high based on this unit price alone. So what is it more likely to rely on to get the volume? I think it is not a single point of explosion, but four lines running together: The catalog structure of programmatic SEO 500+ pages is essentially spreading words. GitHub overflow its footer directly linked to VoltAgent, and the related GitHub warehouse VoltAgent/awesome-design-md has 70,356 stars as of May 4,

  • This magnitude itself will continue to drive developer traffic to the site. The "New Concept Word" card slot is specially made for What is DESIGN.md? This kind of explanation page is obviously grabbing the concept word DESIGN.md, rather than just piling up the brand details page. Social media and circle communication This kind of site is naturally suitable for spreading in X, GitHub, and AI coding agent circles, because it does not sell traditional templates, but a new narrative of "letting AI agents understand the design language." So if you ask me, how is the SEO of this site? My answer is: It’s not the kind of site that “sucks at SEO but got popular by luck”. On the contrary, its technical foundation and structural awareness are online. But if you ask me, has it been publicly verified as a high-traffic, high-income website? Then the answer doesn't sound so good. At least based on public information, we can’t draw such a conclusion. A more accurate statement should be: This is a very smart programmatic SEO + GitHub distribution station. The monetization path has already begun. However, the most common mistake for outsiders now is to directly convert its impressions, collection scale and GitHub popularity into stable income. Between these three things, there is a whole transformation link. If you are also building a content site, tool site or AI directory site, I think the most worthwhile thing to learn from this case is not "how much money it made", but the following: First, put together a structure that search engines can understand, a concept word page, a details page template, a social media communication portal and a charging page. It doesn’t necessarily make you a lot of money right away. But it will at least make your site look like a business, not just a beautiful link. refer to:

Visual summary

Article argument map

Generated from the post's content graph

FORMATTOPICCAPABILITYMARKETcoverscoverscoverscoverssignalssignalssignalssignalsFORMATarticle featureTOPICagentsTOPICtechnical distributionTOPICretrievalTOPICtoolingCAPABILITYagent workflowCAPABILITYdeveloper toolingCAPABILITYevaluationCAPABILITYtechnical writingCAPABILITYproduct surface
Mermaid outline
flowchart LR
  format-article["article feature"]
  topic-agents["agents"]
  topic-technical-distribution["technical distribution"]
  topic-retrieval["retrieval"]
  topic-tooling["tooling"]
  capability-agent-workflow["agent workflow"]
  capability-developer-tooling["developer tooling"]
  capability-evaluation["evaluation"]
  capability-technical-writing["technical writing"]
  capability-product-surface["product surface"]
  format-article -->|covers| topic-agents
  format-article -->|covers| topic-technical-distribution
  format-article -->|covers| topic-retrieval
  format-article -->|covers| topic-tooling
  format-article -->|signals| capability-agent-workflow
  format-article -->|signals| capability-developer-tooling
  format-article -->|signals| capability-evaluation
  format-article -->|signals| capability-technical-writing

Visual structure

Essay structure map

Built from summary and key paragraph positions

Don’t treat impressions as traffic: I looked at getdesign.md and found that it is mor...THESISDon’t treatimpressions astraffic: I looked atgetdesign.md and foundSIGNALDon’t treatimpressions astraffic: I looked atgetdesign.md and foundOPERATOR2026. This magnitudeitself will continueto drive developertraffic to the site.IMPLICATION2026. This magnitudeitself will continueto drive developertraffic to the site.
Mermaid outline
flowchart LR
  thesis["Don’t treat impressions as traffic: I looked at getdesign.md and found that it is more like a programmatic..."]
  signal["Don’t treat impressions as traffic: I looked at getdesign.md and found that it is more like a programmatic..."]
  operator["2026. This magnitude itself will continue to drive developer traffic to the site. The \"New Concept Word\" ca..."]
  implication["2026. This magnitude itself will continue to drive developer traffic to the site. The \"New Concept Word\" ca..."]
  thesis -->|frames| signal
  signal -->|develops| operator
  operator -->|lands in| implication

Source: View the original post