Where is Claude Code demand growing, and which searches are still unclaimed?

1,812,840 searches a month, and not one measured cluster is falling. The fastest-growing part of it — "codex vs claude code", "cursor vs claude code" — is also the easiest to rank for, at a difficulty of 10 against 100 for the product name itself.

claude code · the United States · measured 2026-09-03 · published 2026-09-03

Measured from

Google search demand — monthly search volume, cost-per-click, difficulty and 12 months of history

Google organic results — which domains rank across the measured terms, and where; a domain's own ranking footprint; domains competing across the same keyword set

Google Ads Transparency Center — who is buying ads against these terms; the ads themselves, with the dates they ran

What do the numbers show?

• Every cluster measured here is rising: the name "claude code" is up 127%, installation up 106%, pricing up 90%, skills up 315% and subscription questions up 1074%. Nothing measured is falling.
• The comparison searches are growing faster than the product itself and are the cheapest thing here to rank for: "codex vs claude code" and its cluster grew 219% and carry $692,308 of monthly click value at a difficulty of 10, against 100 for the product name.
• Reddit is the top-ranking domain on 72% of the largest Claude Code terms, at a median position of 3, earning an estimated $97,617 a month of equivalent paid traffic from them.
• claude.com ranks for 12,515 keywords with 3,309 of them at position one, worth an estimated $1,358,671 a month in equivalent paid traffic.
• One domain was found advertising against Claude Code searches — claudecodeads.com — and it is not a competing tool: it sells ad placements inside Claude Code's own output.
• Demand here is overwhelmingly people asking rather than people navigating: 85% of the measured volume carries informational or commercial intent, across 206 terms and 1,812,840 searches a month.

Where is the demand, and which way is it moving?

206 terms carry 1,812,840 searches a month in the United States. Each group below is a set of phrases sharing a modifier — the word that says what the searcher actually wanted — and each is placed by how its own demand has moved against what that demand costs to buy.

The twenty largest phrases, and which way each one moved

Cluster averages say which kind of question is growing. These are the phrases themselves, each with its own direction, so a page written against this category can be aimed at a term rather than at a theme.

A change above a few hundred percent is printed with the two windows behind it, because "a hundred searches became three thousand" and "+2,922%" are the same fact and only one of them can be mistaken for a category doubling.

Change is measured 12-month swing, trailing quarter against the opening quarter, on each cluster's own summed monthly series rather than as an average of its terms — one term going from a handful of searches to a few hundred would otherwise drag a flat cluster into "surging". A cluster marked not rated had too little history, or history covering too little of its volume, to place; that is reported rather than shown as flat.

"claude code" and "claude-code": what are these 3 terms telling you?

The two spellings of the name, "claude code" and "claude-code", draw 550,000 searches a month each and the pair is up 127%. It is also the hardest thing on this map to rank for, at a difficulty of 100. Nobody is taking this position from the vendor; the interesting question is what happens beside it.

"install claude code": what are these 13 terms telling you?

66,760 searches a month on getting it running, up 106%, at a difficulty of 22. "install claude code" alone is 27,100. Installation demand at this size, this late, means the user base is still expanding rather than settling — people do not search for how to install something they already have.

"claude code price" and "pricing": what are these 4 terms telling you?

"claude code price" and "claude code pricing" draw 27,100 searches a month each and the pair is up 90%. Pricing demand rising in step with the product name is a buying signal rather than a churn signal; it is what a market looks like when new people are deciding, not when existing users are leaving.

"codex vs claude code", "cursor vs claude code": what are these 17 terms telling you?

The cluster that matters. "codex vs claude code" draws 14,800 searches a month, "cursor vs claude code" 8,100, and the group grew 219% — faster than the product name itself. It carries $692,308 of monthly click value at a difficulty of 10, which makes it the cheapest valuable position in this measurement. It is cheap because vendors do not write comparison pages naming their rivals, so the demand is answered by whoever will.

"claude code skills": what are these 2 terms telling you?

"claude code skill" and "claude code skills" are 12,100 searches a month each and up 315% — the fastest-growing feature term measured here. A capability that did not exist as a search a year ago now draws as much demand as the pricing page.

"claude code agents": what are these 8 terms telling you?

"claude code agents", "claude code agent" and the team variants draw 22,460 searches a month, up 76%. Agent demand growing more slowly than skills or subscriptions is worth noticing — the composable parts of this product are being searched harder than the autonomous ones.

"claude code subscription": what are these 2 terms telling you?

Up 1074%, the largest percentage on this map, off a base small enough that the two windows are printed beside it. What it says is simpler than the number: the questions moved from what the product does to what it costs to keep.

"how to use claude code": what are these 5 terms telling you?

"how to use claude code" at 8,100 a month leads a cluster up 164%. Instructional demand rising alongside installation is the same signal read twice: this is a product acquiring users, not retaining them.

"everything claude code" — the community repos: what are these 5 terms telling you?

Five terms around "everything claude code" and its GitHub variants, arriving from nothing inside the window. People are searching for a community catalogue of what the product can do, which is a gap in the documentation stated as a search query.

Who is capturing this demand today?

Across the 100 largest terms, this demand is owned. Coverage is the share of those terms a domain ranks for at all; median position is where it sits when it does.

What are the advertisers actually saying?

The ads below ran against this subject on the dates shown. Each image is the creative as it was served, captured from the public ad transparency record and stored here so the picture and the claim beside it stay the same age; the link goes to the record it came from.

• 

  Gabriel Kapuscinski

  claudecodeads.com

  text ad · ran 2026-06-13 to 2026-07-28

  The transparency record for this ad
• 

  Gabriel Kapuscinski

  claudecodeads.com

  text ad · ran 2026-06-14 to 2026-06-27

  The transparency record for this ad

Source: Google Ads Transparency Center, captured 2026-09-03. Ad transparency shows ads that were served; it does not report spend, and an advertiser absent from it is not proven absent from the auction.

How much search does claude.com actually hold?

The domain ranks for 12,515 keywords, 3,309 of them at position one, worth an estimated $1,358,671 a month in equivalent paid traffic. The churn figures are the more interesting half: 7,709 of those rankings are new and 46,856 were lost, which is the footprint of a domain that has recently moved rather than one losing ground — Anthropic's product pages migrated, and the old addresses left the index while the new ones entered it.

Who else ranks for the words claude.com ranks for?

Across the whole keyword set, the domains sharing the most ground with claude.com are not rival products — they are platforms. YouTube shares 11,849 keywords, Reddit 11,490, Medium 8,980, GitHub 7,075. The competitive picture for a developer tool at this stage is not other vendors' marketing pages; it is the places developers write things down, which is the same finding the ranking data gives from the other direction.

Is Claude Code search demand still growing?

Yes, on every axis measured. The product name is up 127% across the window, installation searches up 106%, pricing up 90%, skills up 315%. Not one of the clusters measured here is falling, which is unusual — most subjects have at least one part of their demand draining away.

Which Claude Code keywords are worth writing for?

The comparisons. "codex vs claude code" draws 14,800 searches a month and "cursor vs claude code" 8,100, the cluster grew 219%, and it sits at a difficulty of 10 against 100 for the product name. Vendors do not write pages naming their rivals, so that demand is answered by whoever will.

Who ranks for Claude Code searches?

Reddit, on 72% of the largest terms at a median position of 3. Across claude.com's whole keyword set the domains sharing the most ground are YouTube at 11,849 shared keywords, Reddit at 11,490 and Medium at 8,980 — platforms where developers write things down, not competing vendors' marketing.

Is anyone advertising against Claude Code?

One domain, claudecodeads.com, and it does not sell a competing tool. Its ads offer to place advertising inside Claude Code's own output — "turn Claude Code's thinking line into income". On a product drawing 1,812,840 searches a month, no rival tool was found bidding at all.

How was this measured?

Seven calls to a public data API, listed at the bottom of this page with their arguments. Search volumes, cost-per-click, difficulty and monthly history come from Google search demand; the ranking and competitor picture from Google organic results; the advertising picture and the ad images from the Google Ads Transparency Center. Anyone can run the same calls and get the same shape.

What does this case study not show?

• This measures search demand in the United States, in English. Claude Code has a global user base and this run does not describe it.
• Search demand is not usage. A product can grow in searches because people are adopting it or because they are stuck, and installation and how-to demand rising together is consistent with both.
• The lost-rankings figure for claude.com is consistent with a domain migration rather than a decline, but this run measured one domain at one moment and cannot separate the two on its own.
• Cluster names are this engine's description of a group of phrases, not a taxonomy anybody else uses. The grouping is reproducible from the terms listed under each one.
• Ad transparency reports ads that were served, not spend. An advertiser absent from it is not proven absent from the auction.
• The Searches/mo column is the source's reported monthly figure, a rolling average, while momentum compares three-month means from the same series. For a term that moved sharply inside the window the two differ, which is why both are printed.
• Click value is monthly searches multiplied by published cost-per-click. It prices the attention, and is not revenue anybody earned.
• 1% of the measured volume carried no published cost-per-click and is excluded from every value figure, rather than counted as zero.