← Run this on your site — free
Conversion Audit · your site through your buyers' eyes · model: Claude Sonnet 5
EN·RU

fortem.dev

64

We ran 6 visits through your site — each visitor type enters by its own scenarios. 4 reached their goal.

/
/blog/
/demo/
/book/
/about/
/audit/
/terms/
/guides/
/versus/
/privacy/
PEPlatform Engineer / Fleet Cost Owner
ELEngineering Lead Evaluating Buy vs. Build
EFECS Fargate Practitioner Researching a Specific Problem
search entry · /audit/Platform Engineer / Fleet Cost Owner
direct entry · /Platform Engineer / Fleet Cost Owner
search entry · /versus/Engineering Lead Evaluating Buy vs. Build
direct entry · /about/Engineering Lead Evaluating Buy vs. Build
search entry · /guides/ECS Fargate Practitioner Researching a Specific Problem
direct entry · /blog/ECS Fargate Practitioner Researching a Specific Problem
click a number to see its findings
✉ Email this fix list
CRITICALyou're losing people here /versus/

The 'default alternative' DIY quotes a six-month build cost of $90–110k in loaded labor on the comparison hub, but the About page describes the identical scenario ('rebuilt the same control plane internally... over six months') as costing five hundred thousand dollars. Same claim, same timeframe, 5x apart.

Why it matters — I'm building a buy-vs-build cost case to bring to my team. If Fortem's own site can't agree on its central DIY number, I can't use either figure in a board/eng-leadership pitch — it undermines the whole comparison.

Evidence · verbatim from crawl
Six months of one senior engineer: $90–110k in loaded labor .
CRITICALyou're losing people here /about/

The About page states the same six-month internal-build story costs $500k, directly conflicting with the $90-110k figure used on the /versus/ page for the exact same scenario.

Why it matters — Credibility on cost math is the whole pitch for a 'buy' decision. A 5x internal inconsistency makes me distrust every other number on the site (audit savings, ROI, pricing).

Evidence · verbatim from crawl
eventually hit the same wall and rebuilt the same control plane internally — badly, over six months, for five hundred thousand dollars in engineering time.
MEDIUMquestionable but okay /audit/

The audit's own example report ends with 'Book a call →' as the only next step after showing modeled savings — there's no self-serve signup or checkout to actually deploy the scheduling/control-plane that would cut the idle spend. Identification is self-serve; elimination isn't.

Why it matters — A platform engineer expecting to size and shut off idle spend independently hits a wall: the only route to actually stopping the spend is a sales call, not a button.

Evidence · verbatim from crawl
Modeled savings: ~$2,840/month. Fortem costs $790/mo — break even in ~1 week. Projected net savings: ~$25k/year. Book a call →
MEDIUMquestionable but okay /audit/

If you don't already have Claude Code, Cursor, Gemini, or Codex installed, the site offers no browser-based or hosted alternative to run the audit — the fallback is immediately a sales call.

Why it matters — Some fleet owners manage infra via other tooling and don't run a CLI coding agent daily; for them the 'no-signup, read-only' promise silently collapses into a sales funnel.

Evidence · verbatim from crawl
No CLI agent? Book a 20-min call →
MEDIUMquestionable but okay https://fortem.dev/audit/

The /audit/ page offers a 'Copy Prompt' button but never displays the actual prompt text on the page. A user has to trust the clipboard content without reading it first, unlike the 'Read the skill source' link which does let you inspect the underlying code.

Why it matters — A platform engineer pasting an unseen string into a CLI agent that has shell/file access wants to read it first, not just trust a button label.

Evidence · verbatim from crawl
Copy Prompt + any CLI agent
MEDIUMquestionable but okay https://fortem.dev/audit/

The audit explains its core read-only ECS calls, but also says an optional Cost Explorer call will fire automatically if the credentials happen to allow it — not as an explicit per-run opt-in.

Why it matters — A cost/security owner wants to control exactly which API calls run each time, not have scope depend silently on whatever the ambient IAM role allows.

Evidence · verbatim from crawl
An optional Cost Explorer cross-check ( ce:GetCostAndUsage , ~$0.01/call) runs only if your credentials already allow it.
MEDIUMquestionable but okay /versus/

DIY is explicitly called 'the most common alternative' and the one most teams like mine actually hit, yet it gets one short paragraph with no dedicated 'Read →' article, while Humanitec, Cortex, and Flightcontrol each get full comparison posts.

Why it matters — My decision is buy vs. build, not 'which vendor.' The page gives vendor category-fit clarity but shortchanges the DIY math I most need for an internal build-vs-buy memo.

Evidence · verbatim from crawl
The most common alternative isn't a competing product — it's a platform engineer spending six months building an internal control plane on top of Terraform, Pulumi, or CDK.
MEDIUMquestionable but okay /versus/

The Humanitec teaser promises 'Pricing, features, and fit explained' but no pricing for Humanitec, Cortex, or Flightcontrol appears anywhere on the comparison hub — only Fortem's own $790/mo price and my own AWS bill are quantified.

Why it matters — I need side-by-side cost numbers to justify buy vs. build to my team. A promise of pricing that isn't delivered on the page forces another click-through with no guarantee of finding real numbers.

Evidence · verbatim from crawl
Pricing, features, and fit explained.
MEDIUMquestionable but okay /audit/

To get a concrete cost number for my own fleet (rather than the generic calculator), I have to copy a prompt into a local CLI coding agent and run read-only AWS calls myself — a 15-minute setup task, not an instant comparison.

Why it matters — I'm still in the evaluation/comparison phase — I want a quick number to sanity-check the vendor claims, not a 15-minute local audit before I've even decided this is worth pursuing.

Evidence · verbatim from crawl
Copy one line into your CLI agent. It reads the skill, runs read-only ECS calls, and writes a per-environment cost breakdown to your disk.
MEDIUMquestionable but okay /about/

The about page explicitly states it withholds the founders' names, saying you only get names on a sales call.

Why it matters — I'm trying to verify operator credibility before spending 20 minutes on a call. Without a name I can't check LinkedIn, GitHub, or a talk history — I have to take the claim on faith, which is exactly what I don't want to do at this stage.

Evidence · verbatim from crawl
We keep this page name-free, and we would rather tell you why than pretend it is normal: the engineering founder still operates production infrastructure in a regulated industry, and we separate those worlds deliberately.
MEDIUMquestionable but okay /about/

The headline operator credentials (100+ ECS environments, 85+ financial institutions, 55% cost cut) have no attached company name, case study link, or reference customer.

Why it matters — Big numbers with no way to check them read the same as any generic SaaS marketing copy — the opposite of what this persona came here to distinguish.

Evidence · verbatim from crawl
The engineering founder operates 100+ AWS ECS environments across 4 regions, today, at a fintech serving 85+ global financial institutions.
MEDIUMquestionable but okay https://fortem.dev/guides/

Search query lands on the guides INDEX, not the specific article. Visitor has to scan a list of 9+ teasers to find the one matching their exact question, instead of landing directly on the monitoring guide.

Why it matters — An impatient practitioner wants the answer now; an extra scan-and-click step is friction that can lose a skimmer before they even see the right title.

Evidence · verbatim from crawl
Practical guides for platform engineers operating 10+ ECS environments. No fluff — real configs, real numbers, real tradeoffs.
MEDIUMquestionable but okay https://fortem.dev/blog/

Immediately after the nav bar, before any blog content, headline, or guide list, the page pushes two commercial CTAs.

Why it matters — I'm in research mode, not buying mode. A wall of CTAs before content reads as vendor content-marketing, which undercuts the 'written by engineers who've operated 100+ environments' trust claim right below it.

Evidence · verbatim from crawl
Run Fleet Audit → Book a 20-min call →
MEDIUMquestionable but okay https://fortem.dev/blog/

The blog intro promises operator-written guides but offers no email signup or visible RSS/subscribe link on the page itself — RSS only shows up in the footer of unrelated pages like /book/.

Why it matters — My whole reason for landing here directly was to bookmark/subscribe for future problems. If I can't subscribe from the blog page itself, I have to remember to come back manually — that's how you lose a returning reader.

Evidence · verbatim from crawl
Guides, comparisons, and use cases. Written by engineers who've operated 100+ ECS environments in production.
COSMETICminor /audit/

The $790/mo price point that's used to sell the ROI story ('break even in ~1 week') sits inside a report explicitly labeled as an illustrative mockup whose numbers 'will differ.' It's unclear if the price itself is fixed or just an example.

Why it matters — A cost-owner deciding whether the tool pays for itself wants a firm number, not one presented alongside a disclaimer that everything shown 'will differ.'

Evidence · verbatim from crawl
Example output — your numbers will differ
COSMETICminor https://fortem.dev/audit/

The page offers a link to inspect the actual skill code before running it, which is exactly what a security-conscious engineer wants, but gives no indication of where it goes (GitHub repo? docs page? version?).

Why it matters — Not knowing the destination (repo, commit, hosted doc) before clicking off-site slows down the verification step this persona specifically wants to do.

Evidence · verbatim from crawl
Read the skill source ↗ · No CLI agent? Book a 20-min call →
COSMETICminor https://fortem.dev/guides/

Page promises 'No fluff' but the same marketing CTA block ('Run Fleet Audit →' / 'Book a 20-min call →') appears twice before any guide content, plus a persistent top nav pushing the product on a page meant for technical research.

Why it matters — Practitioner explicitly wants config-level content, not a sales page; repeated CTAs before the list undercut the 'no fluff' promise and read as marketing, which is what this persona actively distrusts.

Evidence · verbatim from crawl
Run Fleet Audit → Book a 20-min call →
COSMETICminor https://fortem.dev/guides/

The 'Read' links for every guide use generic anchor text with no visible URL or preview of depth (code blocks, Terraform snippets), so the visitor can't tell before clicking whether the article is a real technical writeup or another gated funnel.

Why it matters — A skeptical engineer decides in seconds whether to click; opaque 'Read' links add a small trust tax versus, say, showing an anchor snippet or code preview.

Evidence · verbatim from crawl
Read Guide July 2, 2026 · 13 min read AWS ECS Fargate Security
COSMETICminor https://fortem.dev/guides/

The /blog/ and /guides/ pages list the exact same set of guide entries under different URLs and different headers, with no explanation of what's unique to each.

Why it matters — Minor confusion about which URL is the 'real' hub to bookmark — not a blocker, but muddies the site's mental model for a returning visitor.

Evidence · verbatim from crawl
How Do You Scan ECR Images for Vulnerabilities Before They Deploy?
COSMETICminor https://fortem.dev/blog/

With 9+ guides spanning very different ops problems (DR, RBAC, ALB, monitoring, security, SOC2, ECR scanning), there's no tag, category filter, or search box — every visit means scanning the whole list of teasers again.

Why it matters — When I come back with a specific problem (say, RBAC), I want to jump straight there, not re-read titles for DR, SOC2, and ALB to find it.

Evidence · verbatim from crawl
Guides, comparisons, and use cases. Written by engineers who've operated 100+ ECS environments in production.
01

Executive summary

Product delivers on its core promise — read-only, no-signup fleet audits with real IAM call transparency — and two of five personas hit clean 'achieved' outcomes. But the site actively damages its own trust story: the DIY build-cost claim is $90-110k on /versus/ and $500k on /about/, a 5x contradiction that a buy-vs-build lead catches immediately and it poisons every other number on the site. Fix the credibility leak first, then close the two structural gaps: no self-serve path from audit report to actual fix (everything ends at 'Book a call'), and zero competitor pricing despite promising it.

02

Who we tested with

Senior platform/DevOps engineer responsible for 10+ ECS Fargate environments across accounts and regions, tasked with cutting cloud spend without breaking prod.

Intent

Find a way to identify and eliminate idle dev/staging spend without rewriting infra code, and validate the tool is read-only/safe before adopting it.

Why this persona

Homepage headline 'Cut dev/staging costs 60-70% without touching your infra code' and the Fleet Audit's 'Read-only, 15 minutes, nothing leaves your machine' pitch are built for someone who owns ECS spend and needs a safe first step.

search entry · /audit/direct entry · /

Engineering lead or CTO at a 10-50 engineer company running ECS Fargate, evaluating whether to buy a control plane or keep building one internally.

Intent

Decide between building an internal ECS control plane, adopting a general IDP, or buying Fortem — comparing cost, fit, and category before committing engineering time.

Why this persona

The /versus/ page directly targets this decision ('the most common alternative isn't a competing product — it's a platform engineer spending six months building an internal control plane') and the About page's '$500k in engineering time' anecdote is aimed at leaders weighing build cost.

search entry · /versus/direct entry · /about/

Individual contributor backend/infra engineer researching ECS Fargate operational tasks (monitoring, DR, RBAC, ALB setup) who hits Fortem's technical guides while solving a specific problem.

Intent

Solve a concrete ECS Fargate operational problem (monitoring, security, DR, image scanning) using a detailed guide, and discover Fortem as a byproduct.

Why this persona

The /guides/ and /blog/ pages are packed with narrow technical how-tos ('How Do You Monitor ECS Fargate Across 10+ Environments?', 'AWS ECS Fargate Security: What You Actually Configure') written to capture engineers Googling operational pain points.

search entry · /guides/direct entry · /blog/
03

Verdicts by persona

✓ achieved — got what they came for · ◐ partial — got some of it · ✕ failed — left with nothing
6 journeys → 5 made it past the entry page → 4 converted

Got the no-signup, read-only fleet audit exactly as promised — no sales call needed to see the numbers.

  • Entry page matches the search intent word-for-word: read-only, no login, dollar figures shown upfront.
  • Explicit IAM call list makes the 'safe to run' claim verifiable, not just asserted.
  • But actually cutting the idle spend afterward funnels straight into 'Book a call' — no self-serve deploy/checkout.
  • No CLI agent installed? Then the 'no-signup' path quietly becomes a sales call too.

Found the exact audit prompt and read-only proof — ready to run it in Claude Code.

  • Homepage promise ('read-only, 15 min, nothing leaves your machine') is repeated almost verbatim on /audit/ — strong message match.
  • Audit page lists the exact IAM read calls, letting an engineer verify scope before running anything.
  • 'Read the skill source' link lets a skeptical engineer audit the code — but the prompt text itself isn't shown before copying.
  • Optional Cost Explorer call auto-fires if permissions allow it, not a clear explicit opt-in each run.

Got category clarity, not a clean buy-vs-build number to take to my team.

  • Versus hub nails category positioning for Humanitec, Cortex, and Flightcontrol in one glance.
  • Same DIY six-month build claim costs $90-110k on one page and $500k on another — trust broken.
  • No competitor pricing anywhere; the promised 'pricing explained' never shows up.
  • Ends up needing to book a call just to finish the cost math I came to do myself.

Do this →Reconcile the DIY six-month cost figure across /about/ and /versus/ (pick one number and cite the same math everywhere), and add an actual pricing/feature comparison table for Humanitec, Cortex, and Flightcontrol to the versus hub instead of prose teasers.

Lead gets real operator proof and a direct buy-vs-build breakdown, books the call.

  • About page delivers concrete operator numbers — 100+ envs, 55% cost cut, org-wide IDP adoption.
  • Versus page directly answers the build-vs-IDP-vs-buy question with real cost math ($90-110k DIY build).
  • Names withheld until the call undercuts the 'built by operators, not vendors' claim right when verification matters most.
  • Pricing shown only as a single 'starts at' figure — no sense of how cost scales at 20+ environments.

Found the exact-match guide title, but no visibility into whether the article actually delivers the config.

  • Title and teaser for the monitoring guide match the search query almost word-for-word — strong signal, not generic marketing copy.
  • Guides index itself is cluttered with repeated 'Book a call' / 'Run Fleet Audit' CTAs despite claiming 'No fluff.'
  • Query lands on the index page, forcing a scan-and-click instead of a direct hit on the matching article.
  • Can't confirm from this page whether the linked article actually contains the promised Terraform for_each pattern and Container Insights cost breakdown — the payoff is still one unverified click away.

Do this →Make the /guides/ page (or better, the specific monitoring article) the actual SEO landing target for this query, and strip the duplicate CTA blocks from the guide index so the technical list is the first thing a searcher sees.

Found the exact operational guide needed; Fortem itself surfaced as a bonus.

  • Guide titles matched real ops problems almost word-for-word — DR, RBAC, ALB, monitoring, ECR scanning.
  • Teaser copy was specific enough (Terraform patterns, real cost numbers) to feel credible, not marketing fluff.
  • Sales CTAs up front on the blog page feel pushy for a research-mode visit.
  • No subscribe/RSS option on the blog page itself, despite wanting to bookmark for future problems.
04

Key findings

IP-01 CRITICALyou're losing people here /versus/

DIY cost figure contradicts the About page by 5x

The 'default alternative' DIY quotes a six-month build cost of $90–110k in loaded labor on the comparison hub, but the About page describes the identical scenario ('rebuilt the same control plane internally... over six months') as costing five hundred thousand dollars. Same claim, same timeframe, 5x apart.

Why it matters — I'm building a buy-vs-build cost case to bring to my team. If Fortem's own site can't agree on its central DIY number, I can't use either figure in a board/eng-leadership pitch — it undermines the whole comparison.

Evidence · verbatim from crawl
Six months of one senior engineer: $90–110k in loaded labor .
IP-02 CRITICALyou're losing people here /about/

About page's $500k DIY claim contradicts versus hub's $90-110k

The About page states the same six-month internal-build story costs $500k, directly conflicting with the $90-110k figure used on the /versus/ page for the exact same scenario.

Why it matters — Credibility on cost math is the whole pitch for a 'buy' decision. A 5x internal inconsistency makes me distrust every other number on the site (audit savings, ROI, pricing).

Evidence · verbatim from crawl
eventually hit the same wall and rebuilt the same control plane internally — badly, over six months, for five hundred thousand dollars in engineering time.
IP-03 MEDIUMquestionable but okay /audit/

No self-serve path from report to actual fix

The audit's own example report ends with 'Book a call →' as the only next step after showing modeled savings — there's no self-serve signup or checkout to actually deploy the scheduling/control-plane that would cut the idle spend. Identification is self-serve; elimination isn't.

Why it matters — A platform engineer expecting to size and shut off idle spend independently hits a wall: the only route to actually stopping the spend is a sales call, not a button.

Evidence · verbatim from crawl
Modeled savings: ~$2,840/month. Fortem costs $790/mo — break even in ~1 week. Projected net savings: ~$25k/year. Book a call →
IP-04 MEDIUMquestionable but okay /audit/

No CLI agent = forced into a sales call

If you don't already have Claude Code, Cursor, Gemini, or Codex installed, the site offers no browser-based or hosted alternative to run the audit — the fallback is immediately a sales call.

Why it matters — Some fleet owners manage infra via other tooling and don't run a CLI coding agent daily; for them the 'no-signup, read-only' promise silently collapses into a sales funnel.

Evidence · verbatim from crawl
No CLI agent? Book a 20-min call →
IP-05 MEDIUMquestionable but okay https://fortem.dev/audit/

Prompt text hidden behind a copy button

The /audit/ page offers a 'Copy Prompt' button but never displays the actual prompt text on the page. A user has to trust the clipboard content without reading it first, unlike the 'Read the skill source' link which does let you inspect the underlying code.

Why it matters — A platform engineer pasting an unseen string into a CLI agent that has shell/file access wants to read it first, not just trust a button label.

Evidence · verbatim from crawl
Copy Prompt + any CLI agent
05

Remediation — quick wins

  1. Reconcile the DIY six-month build cost — pick one number and use it identically on /about/ and /versus/ (currently $90-110k vs $500k for the same claim).
  2. Strip the 'Book a call' CTAs and duplicate teasers off /guides/ and /blog/ — a page that promises 'no fluff' shouldn't lead with two sales CTAs before any content.
  3. Add a subscribe/RSS link directly on /blog/ — researchers who land here explicitly want to bookmark for later and currently have no way to.
06

Method

Each persona/scenario was simulated as a fresh, isolated model invocation whose context contained only the persona definition, the entry scenario and the crawled page text — zero prior knowledge of the site, no shared state between walkthroughs. Every finding carries a verbatim evidence quote from the crawl; findings whose quotes could not be verified against the crawled text were automatically discarded (1 discarded in this run).
07

Walkthrough transcripts

✓ ACHIEVED Platform Engineer / Fleet Cost Owner search entry · /audit/ · findings: 3

Senior platform/DevOps engineer responsible for 10+ ECS Fargate environments across accounts and regions, tasked with cutting cloud spend without breaking prod.
Find a way to identify and eliminate idle dev/staging spend without rewriting infra code, and validate the tool is read-only/safe before adopting it.

QUERY “reduce ECS Fargate dev staging costs without changing infra code”
EXPECTS A no-signup, read-only way to map ECS environments and see exact idle-cost dollar figures before any sales conversation.

Got the no-signup, read-only fleet audit exactly as promised — no sales call needed to see the numbers.

  1. /audit/
    Lands on /audit/ from the search query. Headline '1 prompt · read-only · 0 bytes uploaded' and title 'Map your ECS fleet, see your savings' immediately match the search intent — no scrolling needed to confirm relevance.
  2. /audit/
    Reads the tagline: 'Copy one line into your CLI agent. It reads the skill, runs read-only ECS calls, and writes a per-environment cost breakdown to your disk. Nothing is transmitted — no login, no AWS credentials for us.' This directly answers 'no-signup, read-only' — decides to keep reading instead of bouncing.
  3. /audit/
    Scrolls to 'Exactly what you'll get' and reviews the example HTML report — sees concrete per-environment dollar figures ($4,200/mo → $1,360 with scheduling, −$2,840/mo savings). This is the 'exact idle-cost dollar figures' the persona came for, and it's shown before any sales contact.
  4. /audit/
    Reads 'Your data never leaves your machine' and the specific read-only IAM call list (ecs:ListClusters, DescribeClusters, ListServices, DescribeServices...). This is the safety validation the persona needs before running anything against production AWS accounts — treats it as credible because it names exact API calls rather than just saying 'safe.'
  5. /audit/
    Notices 'Read the skill source ↗' link — the option to inspect the actual prompt/code before running it satisfies the last bit of due diligence a senior engineer would want; doesn't need to click through since the permission list already answers the core question.
  6. /audit/
    Clicks 'Copy Prompt + any CLI agent' to grab the one-liner, intending to run it locally in an existing CLI agent — no signup, no call, no credentials handed over. Primary intent satisfied on the entry page itself.

Findings

IP-03 MEDIUMquestionable but okay /audit/

No self-serve path from report to actual fix

The audit's own example report ends with 'Book a call →' as the only next step after showing modeled savings — there's no self-serve signup or checkout to actually deploy the scheduling/control-plane that would cut the idle spend. Identification is self-serve; elimination isn't.

Why it matters — A platform engineer expecting to size and shut off idle spend independently hits a wall: the only route to actually stopping the spend is a sales call, not a button.

Evidence · verbatim from crawl
Modeled savings: ~$2,840/month. Fortem costs $790/mo — break even in ~1 week. Projected net savings: ~$25k/year. Book a call →
IP-04 MEDIUMquestionable but okay /audit/

No CLI agent = forced into a sales call

If you don't already have Claude Code, Cursor, Gemini, or Codex installed, the site offers no browser-based or hosted alternative to run the audit — the fallback is immediately a sales call.

Why it matters — Some fleet owners manage infra via other tooling and don't run a CLI coding agent daily; for them the 'no-signup, read-only' promise silently collapses into a sales funnel.

Evidence · verbatim from crawl
No CLI agent? Book a 20-min call →
IP-15 COSMETICminor /audit/

Pricing figure buried inside a hedged mockup

The $790/mo price point that's used to sell the ROI story ('break even in ~1 week') sits inside a report explicitly labeled as an illustrative mockup whose numbers 'will differ.' It's unclear if the price itself is fixed or just an example.

Why it matters — A cost-owner deciding whether the tool pays for itself wants a firm number, not one presented alongside a disclaimer that everything shown 'will differ.'

Evidence · verbatim from crawl
Example output — your numbers will differ
✓ ACHIEVED Platform Engineer / Fleet Cost Owner direct entry · / · findings: 3

Senior platform/DevOps engineer responsible for 10+ ECS Fargate environments across accounts and regions, tasked with cutting cloud spend without breaking prod.
Find a way to identify and eliminate idle dev/staging spend without rewriting infra code, and validate the tool is read-only/safe before adopting it.

EXPECTS A quick way to run the fleet audit prompt in Claude Code/Cursor and see savings, since they came back after hearing about it.

Found the exact audit prompt and read-only proof — ready to run it in Claude Code.

  1. https://fortem.dev/
    Land directly on the homepage. Hero line says 'Cut dev/staging costs 60–70% without touching your infra code' and 'Copy one line into Claude Code or any CLI agent... Read-only, 15 minutes, nothing leaves your machine.' This is exactly the pitch I remembered — a CLI-agent fleet audit, no infra rewrite, read-only. I click 'Run Fleet Audit →'.
  2. https://fortem.dev/audit/
    Land on /audit/. Tagline confirms scope: '1 prompt · read-only · 0 bytes uploaded'. Scroll past the copy-prompt CTA to 'Your data never leaves your machine' — it lists the exact IAM calls used (ecs:ListClusters, ecs:DescribeClusters, ecs:ListServices, ecs:DescribeServices...) plus the caveat that an optional Cost Explorer check runs if my creds allow it. I also see 'Read the skill source ↗' to verify the code myself before running anything. The example output section shows the exact three files I'd get (HTML report, YAML config, DIY CloudFormation scheduler) with realistic per-env cost numbers. This matches what I came for — I click 'Copy Prompt' and plan to check the skill source next before pasting into Claude Code.

Findings

IP-05 MEDIUMquestionable but okay https://fortem.dev/audit/

Prompt text hidden behind a copy button

The /audit/ page offers a 'Copy Prompt' button but never displays the actual prompt text on the page. A user has to trust the clipboard content without reading it first, unlike the 'Read the skill source' link which does let you inspect the underlying code.

Why it matters — A platform engineer pasting an unseen string into a CLI agent that has shell/file access wants to read it first, not just trust a button label.

Evidence · verbatim from crawl
Copy Prompt + any CLI agent
IP-06 MEDIUMquestionable but okay https://fortem.dev/audit/

Cost Explorer call auto-fires based on IAM permissions

The audit explains its core read-only ECS calls, but also says an optional Cost Explorer call will fire automatically if the credentials happen to allow it — not as an explicit per-run opt-in.

Why it matters — A cost/security owner wants to control exactly which API calls run each time, not have scope depend silently on whatever the ambient IAM role allows.

Evidence · verbatim from crawl
An optional Cost Explorer cross-check ( ce:GetCostAndUsage , ~$0.01/call) runs only if your credentials already allow it.
IP-16 COSMETICminor https://fortem.dev/audit/

'Read the skill source' link has no context

The page offers a link to inspect the actual skill code before running it, which is exactly what a security-conscious engineer wants, but gives no indication of where it goes (GitHub repo? docs page? version?).

Why it matters — Not knowing the destination (repo, commit, hosted doc) before clicking off-site slows down the verification step this persona specifically wants to do.

Evidence · verbatim from crawl
Read the skill source ↗ · No CLI agent? Book a 20-min call →
◐ PARTIAL Engineering Lead Evaluating Buy vs. Build search entry · /versus/ · findings: 5

Engineering lead or CTO at a 10-50 engineer company running ECS Fargate, evaluating whether to buy a control plane or keep building one internally.
Decide between building an internal ECS control plane, adopting a general IDP, or buying Fortem — comparing cost, fit, and category before committing engineering time.

QUERY “Fortem vs Humanitec ECS fleet operations comparison”
EXPECTS Honest, specific comparisons against Flightcontrol, Humanitec, Cortex, and DIY so they can justify a buy decision to their team.

Got category clarity, not a clean buy-vs-build number to take to my team.

  1. /versus/
    Arrive from a search for 'Fortem vs Humanitec ECS fleet operations comparison' directly on /versus/. Scan the card list — immediately spot 'Fortem vs Humanitec: ECS Fleet Operations vs General-Purpose IDP' with the line 'Humanitec's Container Driver explicitly excludes ECS... you're comparing the wrong category of tool.' That answers my core category question without even clicking through. I also clock the Cortex teaser ('Fortem operates your ECS Fargate fleet specifically... which one you need — and when to use both') and the Flightcontrol note ('If you're running 1–3 ECS services, you should probably use Flightcontrol') — three of my four comparisons framed at a glance.
  2. /versus/
    Scroll to the bottom section, 'The default alternative: DIY' — this is the real competitor I'm weighing internally. It gives a number: 'Six months of one senior engineer: $90–110k in loaded labor.' Useful — I note it down to bring to my team, but the section is thin, just one paragraph, no dedicated article like the vendors get.
  3. /about/
    Click 'About' in the footer to check if these people actually operate ECS at scale before I trust their math. Good operator credentials (100+ envs, fintech, SOC2). But then I hit: 'rebuilt the same control plane internally — badly, over six months, for five hundred thousand dollars in engineering time.' That's the same six-month DIY story I just read on /versus/ — quoted at $90-110k there and $500k here. I stop trusting either number.
  4. /audit/
    Try 'Run Fleet Audit' hoping for a fast, independent number to replace the shaky DIY estimate. Turns out it requires copying a prompt into my own CLI coding agent and running read-only AWS calls locally — a 15-minute task, not the instant sanity-check I wanted mid-comparison.
  5. /book/
    No competitor pricing was ever shown for Humanitec, Cortex, or Flightcontrol — only Fortem's $790/mo and my own AWS bill got numbers. Land on 'Book a call' as the fallback, since finishing this comparison myself with the numbers on the site isn't possible.

Findings

IP-01 CRITICALyou're losing people here /versus/

DIY cost figure contradicts the About page by 5x

The 'default alternative' DIY quotes a six-month build cost of $90–110k in loaded labor on the comparison hub, but the About page describes the identical scenario ('rebuilt the same control plane internally... over six months') as costing five hundred thousand dollars. Same claim, same timeframe, 5x apart.

Why it matters — I'm building a buy-vs-build cost case to bring to my team. If Fortem's own site can't agree on its central DIY number, I can't use either figure in a board/eng-leadership pitch — it undermines the whole comparison.

Evidence · verbatim from crawl
Six months of one senior engineer: $90–110k in loaded labor .
IP-02 CRITICALyou're losing people here /about/

About page's $500k DIY claim contradicts versus hub's $90-110k

The About page states the same six-month internal-build story costs $500k, directly conflicting with the $90-110k figure used on the /versus/ page for the exact same scenario.

Why it matters — Credibility on cost math is the whole pitch for a 'buy' decision. A 5x internal inconsistency makes me distrust every other number on the site (audit savings, ROI, pricing).

Evidence · verbatim from crawl
eventually hit the same wall and rebuilt the same control plane internally — badly, over six months, for five hundred thousand dollars in engineering time.
IP-07 MEDIUMquestionable but okay /versus/

DIY — the actual alternative I'm weighing — gets no real depth

DIY is explicitly called 'the most common alternative' and the one most teams like mine actually hit, yet it gets one short paragraph with no dedicated 'Read →' article, while Humanitec, Cortex, and Flightcontrol each get full comparison posts.

Why it matters — My decision is buy vs. build, not 'which vendor.' The page gives vendor category-fit clarity but shortchanges the DIY math I most need for an internal build-vs-buy memo.

Evidence · verbatim from crawl
The most common alternative isn't a competing product — it's a platform engineer spending six months building an internal control plane on top of Terraform, Pulumi, or CDK.
IP-08 MEDIUMquestionable but okay /versus/

No competitor pricing shown despite promising 'pricing explained'

The Humanitec teaser promises 'Pricing, features, and fit explained' but no pricing for Humanitec, Cortex, or Flightcontrol appears anywhere on the comparison hub — only Fortem's own $790/mo price and my own AWS bill are quantified.

Why it matters — I need side-by-side cost numbers to justify buy vs. build to my team. A promise of pricing that isn't delivered on the page forces another click-through with no guarantee of finding real numbers.

Evidence · verbatim from crawl
Pricing, features, and fit explained.
IP-09 MEDIUMquestionable but okay /audit/

Getting a real fleet number requires running a local tool, not instant

To get a concrete cost number for my own fleet (rather than the generic calculator), I have to copy a prompt into a local CLI coding agent and run read-only AWS calls myself — a 15-minute setup task, not an instant comparison.

Why it matters — I'm still in the evaluation/comparison phase — I want a quick number to sanity-check the vendor claims, not a 15-minute local audit before I've even decided this is worth pursuing.

Evidence · verbatim from crawl
Copy one line into your CLI agent. It reads the skill, runs read-only ECS calls, and writes a per-environment cost breakdown to your disk.
✓ ACHIEVED Engineering Lead Evaluating Buy vs. Build direct entry · /about/ · findings: 2

Engineering lead or CTO at a 10-50 engineer company running ECS Fargate, evaluating whether to buy a control plane or keep building one internally.
Decide between building an internal ECS control plane, adopting a general IDP, or buying Fortem — comparing cost, fit, and category before committing engineering time.

EXPECTS Proof the vendor is run by operators who've actually run ECS at scale, not a generic SaaS pitch, before booking a call.

Lead gets real operator proof and a direct buy-vs-build breakdown, books the call.

  1. /about/
    Land on /about/ looking for proof this is run by real ECS operators, not a generic vendor. Find concrete operator stats — 100+ environments, 85+ financial institutions, 55% cost cut, an internal IDP built and adopted org-wide — plus SOC2/HIPAA/ISO experience. Good signal, but the page is deliberately name-free, so I can't independently verify any of it yet. Click the logo/breadcrumb 'Home' to see the actual product before I judge further.
  2. https://fortem.dev
    On the homepage I see the live-looking fleet console, a cost calculator, and a concrete pricing anchor ('Fortem starts at $790/month'). This grounds the earlier claims in something tangible — a real UI, real math. I still need the buy-vs-build-vs-IDP comparison explicitly, so I follow the 'Compare' footer links.
  3. /versus/
    Land on /versus/ — this page answers my actual question directly: it stacks Fortem against Humanitec (general IDP, explicitly wrong category for ECS) and against DIY build ('$90–110k in loaded labor' for six months, then you own and maintain it forever). This is the exact comparison I came here to make, with real numbers instead of vague positioning.
  4. /audit/
    Curious whether I can validate the savings claim without giving up any access, I check /audit/. It's a read-only CLI prompt, no AWS credentials, nothing uploaded — low-risk way to test the pitch against my own fleet before committing to anything. This lowers the bar to engage further.
  5. /book/
    Decide the operator story plus the explicit build-vs-buy-vs-IDP math is enough to justify a 20-minute call with the person who actually built the IAM model, rather than a sales rep. Click through to /book/.

Findings

IP-10 MEDIUMquestionable but okay /about/

Founders anonymized until you book a call

The about page explicitly states it withholds the founders' names, saying you only get names on a sales call.

Why it matters — I'm trying to verify operator credibility before spending 20 minutes on a call. Without a name I can't check LinkedIn, GitHub, or a talk history — I have to take the claim on faith, which is exactly what I don't want to do at this stage.

Evidence · verbatim from crawl
We keep this page name-free, and we would rather tell you why than pretend it is normal: the engineering founder still operates production infrastructure in a regulated industry, and we separate those worlds deliberately.
IP-11 MEDIUMquestionable but okay /about/

Operator claims are unverifiable as written

The headline operator credentials (100+ ECS environments, 85+ financial institutions, 55% cost cut) have no attached company name, case study link, or reference customer.

Why it matters — Big numbers with no way to check them read the same as any generic SaaS marketing copy — the opposite of what this persona came here to distinguish.

Evidence · verbatim from crawl
The engineering founder operates 100+ AWS ECS environments across 4 regions, today, at a fintech serving 85+ global financial institutions.
◐ PARTIAL ECS Fargate Practitioner Researching a Specific Problem search entry · /guides/ · findings: 3

Individual contributor backend/infra engineer researching ECS Fargate operational tasks (monitoring, DR, RBAC, ALB setup) who hits Fortem's technical guides while solving a specific problem.
Solve a concrete ECS Fargate operational problem (monitoring, security, DR, image scanning) using a detailed guide, and discover Fortem as a byproduct.

QUERY “how do you monitor ECS Fargate across multiple environments”
EXPECTS A practical, config-level answer to an ECS Fargate operational question, not a marketing page.

Found the exact-match guide title, but no visibility into whether the article actually delivers the config.

  1. https://fortem.dev/guides/
    Arrives from the search query 'how do you monitor ECS Fargate across multiple environments' directly on the guides index (not the homepage — good, no upfront pricing/demo noise). Ignores the top nav and the repeated 'Run Fleet Audit →' / 'Book a 20-min call →' CTAs. Skims the list of guide teasers looking for a title that matches the question.
  2. https://fortem.dev/guides/
    Spots the third entry: 'How Do You Monitor ECS Fargate Across 10+ Environments?' with teaser 'Which metrics to alarm on, what Container Insights actually costs per metric at fleet scale, and a Terraform for_each pattern that alarms your whole fleet from one module.' This is a near-exact match for the search intent — specific, technical, no vague marketing language. Decides this is the right resource and clicks 'Read' expecting the actual metrics list and Terraform config.

Findings

IP-12 MEDIUMquestionable but okay https://fortem.dev/guides/

Search lands on guide index, not the matching article

Search query lands on the guides INDEX, not the specific article. Visitor has to scan a list of 9+ teasers to find the one matching their exact question, instead of landing directly on the monitoring guide.

Why it matters — An impatient practitioner wants the answer now; an extra scan-and-click step is friction that can lose a skimmer before they even see the right title.

Evidence · verbatim from crawl
Practical guides for platform engineers operating 10+ ECS environments. No fluff — real configs, real numbers, real tradeoffs.
IP-17 COSMETICminor https://fortem.dev/guides/

Marketing CTAs clutter a page that claims 'no fluff'

Page promises 'No fluff' but the same marketing CTA block ('Run Fleet Audit →' / 'Book a 20-min call →') appears twice before any guide content, plus a persistent top nav pushing the product on a page meant for technical research.

Why it matters — Practitioner explicitly wants config-level content, not a sales page; repeated CTAs before the list undercut the 'no fluff' promise and read as marketing, which is what this persona actively distrusts.

Evidence · verbatim from crawl
Run Fleet Audit → Book a 20-min call →
IP-18 COSMETICminor https://fortem.dev/guides/

'Read' links give no preview of destination depth

The 'Read' links for every guide use generic anchor text with no visible URL or preview of depth (code blocks, Terraform snippets), so the visitor can't tell before clicking whether the article is a real technical writeup or another gated funnel.

Why it matters — A skeptical engineer decides in seconds whether to click; opaque 'Read' links add a small trust tax versus, say, showing an anchor snippet or code preview.

Evidence · verbatim from crawl
Read Guide July 2, 2026 · 13 min read AWS ECS Fargate Security
✓ ACHIEVED ECS Fargate Practitioner Researching a Specific Problem direct entry · /blog/ · findings: 4

Individual contributor backend/infra engineer researching ECS Fargate operational tasks (monitoring, DR, RBAC, ALB setup) who hits Fortem's technical guides while solving a specific problem.
Solve a concrete ECS Fargate operational problem (monitoring, security, DR, image scanning) using a detailed guide, and discover Fortem as a byproduct.

EXPECTS More guides written by people who actually operate ECS fleets, to bookmark or subscribe to for future problems.

Found the exact operational guide needed; Fortem itself surfaced as a bonus.

  1. https://fortem.dev/blog/
    Land directly on /blog/ expecting practitioner-written ECS Fargate guides to bookmark. Immediately see the tagline 'Written by engineers who've operated 100+ ECS environments in production' — matches my expectation exactly. Scan the list: titles are phrased as the exact ops questions I have — ECR scanning, cross-account DR, ALB setup, monitoring 10+ environments, security config, RBAC, SOC2 audit prep, multi-account management. This is unusually specific for a vendor blog. Notice the CTA block ('Run Fleet Audit → Book a 20-min call →') sitting right under the nav before any content — a bit salesy, but I ignore it and keep scanning titles. My current problem is fleet monitoring, so I go for that one.
  2. Guide: How Do You Monitor ECS Fargate Across 10+ Environments?
    Click 'Read →' on 'How Do You Monitor ECS Fargate Across 10+ Environments?'. The teaser already promised concrete specifics — which metrics to alarm on, real Container Insights cost-per-metric numbers, and a Terraform for_each pattern to alarm the whole fleet from one module — not vague fluff. That level of specificity is what I came for. I also notice the same persistent CTAs and Fortem branding surrounding the article, which is how the product enters my radar as a byproduct of reading the guide.
  3. https://fortem.dev
    Curious what this Fortem thing actually is, click through to the homepage. See 'Cut dev/staging costs 60–70% without touching your infra code' plus a live demo table of environments and a cost calculator. I recognize this as a real ops tool (schedule dev/staging down, AI diagnoses crashed tasks) relevant to my day job managing a fleet — worth remembering, not something I need right now.
  4. https://fortem.dev/about/
    Check the About page out of habit before trusting the guide I just read. 'Built by operators, not vendors' — the founder 'operates 100+ AWS ECS environments across 4 regions, today.' That backs up the credibility of the guide content. Satisfied — I got my answer, and I'm bookmarking the domain for future problems.

Findings

IP-13 MEDIUMquestionable but okay https://fortem.dev/blog/

Sales CTAs crowd blog page before any content

Immediately after the nav bar, before any blog content, headline, or guide list, the page pushes two commercial CTAs.

Why it matters — I'm in research mode, not buying mode. A wall of CTAs before content reads as vendor content-marketing, which undercuts the 'written by engineers who've operated 100+ environments' trust claim right below it.

Evidence · verbatim from crawl
Run Fleet Audit → Book a 20-min call →
IP-14 MEDIUMquestionable but okay https://fortem.dev/blog/

No subscribe/RSS option visible on blog listing

The blog intro promises operator-written guides but offers no email signup or visible RSS/subscribe link on the page itself — RSS only shows up in the footer of unrelated pages like /book/.

Why it matters — My whole reason for landing here directly was to bookmark/subscribe for future problems. If I can't subscribe from the blog page itself, I have to remember to come back manually — that's how you lose a returning reader.

Evidence · verbatim from crawl
Guides, comparisons, and use cases. Written by engineers who've operated 100+ ECS environments in production.
IP-19 COSMETICminor https://fortem.dev/guides/

Blog and Guides pages duplicate the same list

The /blog/ and /guides/ pages list the exact same set of guide entries under different URLs and different headers, with no explanation of what's unique to each.

Why it matters — Minor confusion about which URL is the 'real' hub to bookmark — not a blocker, but muddies the site's mental model for a returning visitor.

Evidence · verbatim from crawl
How Do You Scan ECR Images for Vulnerabilities Before They Deploy?
IP-20 COSMETICminor https://fortem.dev/blog/

No topic filter on a long, growing guide list

With 9+ guides spanning very different ops problems (DR, RBAC, ALB, monitoring, security, SOC2, ECR scanning), there's no tag, category filter, or search box — every visit means scanning the whole list of teasers again.

Why it matters — When I come back with a specific problem (say, RBAC), I want to jump straight there, not re-read titles for DR, SOC2, and ALB to find it.

Evidence · verbatim from crawl
Guides, comparisons, and use cases. Written by engineers who've operated 100+ ECS environments in production.
08

Scope — crawl inventory

10 pages crawled (text-level, no JS rendering) — expand
https://fortem.dev
https://fortem.dev/blog/
https://fortem.dev/demo/
https://fortem.dev/book/
https://fortem.dev/about/
https://fortem.dev/audit/
https://fortem.dev/terms/
https://fortem.dev/guides/
https://fortem.dev/versus/
https://fortem.dev/privacy/