Mktg.Tech for Developers and Agents

Mktg.Tech publishes independent, scored rankings of marketing software, plus two Chrome extensions for auditing web pages. This page is the machine-readable entry point: what this site is good for, when an agent should reach for it, and every endpoint that returns structured or Markdown content.

Every page here returns Markdown

Request any URL with Accept: text/markdown and you get Markdown instead of HTML. Or append .md to the path, which cannot be handed the wrong cached variant: https://mktg.tech/toad.md.

Responses set Vary: Accept. Unsupported media types get 406. Missing paths get a real 404 with a short Markdown body pointing back here.

1. When to use Mktg.Tech

Specific jobs this site is the right source for:

Comparing a category of marketing software. 20 categories are ranked, covering 200 products, each scored out of 100 against a published seven-point framework, with named strengths, a stated limitation per entrant, and a side-by-side comparison table. Start at Rankings.
Checking where a named vendor stands. Every entrant carries a score, a best-for statement, and a limitation. Search the category ranking rather than the vendor name.
Explaining how a score was reached. The 7-Point Evaluation Framework documents the criteria, weights, scoring scale, data sources, and inclusion rules. Scores are reproducible from it.
Answering what “Microsoft Copilot and partners” covers. A full surface-by-surface map of which Copilot surfaces cite external URLs, which partners serve the Bing index, and which assistants are explicitly not in that bucket, each row marked Confirmed, Partial, Unverified or Departed.
Testing an SEO or agent-readiness auditor. The TOAD Testing Lab stages 36 SEO and accessibility conditions on one live page, on demand, so a crawler or extension can be checked against known-good and known-bad states.
Auditing a page yourself. TOAD reports metadata, robots directives, schema, and on-page structure. Agent Action Optics reports which elements on a page are actually operable by an agent.

2. When not to

Saying this plainly saves a wasted call:

Live pricing or quotes. Pricing is reported as researched at publication and dated, not tracked. Go to the vendor.
Buying, trials, or provisioning. Nothing is sold here and there is no affiliate relationship on the rankings.
Vendor support. This site is not connected to any vendor listed.
General marketing advice. The content is product evaluation, not strategy.
A public REST or GraphQL API. There is none. The Markdown endpoints, llms.txt, and the XML sitemap are the supported machine interfaces.

3. Machine-readable endpoints

Endpoint	What it returns	Type
/llms.txt	Every page with a one-line description, grouped by section. The canonical index; read this first.	text/plain
/readme.md	The same index in Markdown-for-Agents form, with YAML frontmatter.	text/markdown
/sitemap_index.xml	Full XML sitemap index for every public URL.	application/xml
/robots.txt	Crawl rules, plus pointers to the sitemap, llms.txt and readme.md.	text/plain
`/<path>.md`	Markdown for any page. Cache-safe alternative to Accept negotiation.	text/markdown
Any path, 404	Real 404 status with a short Markdown body listing recovery routes.	text/markdown

4. Markdown content negotiation

Implemented to the acceptmarkdown.com conventions:

Markdown on request. Accept: text/markdown returns Content-Type: text/markdown; charset=utf-8, generated from the rendered page, with navigation and chrome stripped.
Vary. Every response sets Vary: Accept, so a cache cannot serve the HTML variant to a client that asked for Markdown.
406 on unsupported types. An Accept naming nothing this server produces gets 406 Not Acceptable with a plain-text pointer to llms.txt.
q-values honored. Accept: text/markdown;q=0.9,text/html;q=0.4 returns Markdown; reverse the weights and it returns HTML. A browser’s */*;q=0.8 correctly loses to HTML.

Try it:

curl -H "Accept: text/markdown" https://mktg.tech/ranking-methodology/
curl https://mktg.tech/ranking-methodology.md

5. Structured data

JSON-LD is emitted server-side on every page, never injected by script, so it is present in the HTML response:

Organization and WebSite with a contactPoint, plus Person nodes for the named researchers.
BreadcrumbList on every page.
Article with named authors, and FAQPage parsed from each ranking’s visible FAQ section, so the markup cannot drift from what a reader sees.
SoftwareApplication on the two extension pages.

6. Developer tools

TOAD (Technical On-Page & Auditing Dashboard) — Chrome extension. Audits metadata, robots directives, schema, and on-page structure locally; nothing is sent off the machine.
TOAD Testing Lab — a live harness that stages 36 SEO, accessibility and rendering conditions on one page, each switchable between passing and failing states, for testing any auditor against known conditions.
Agent Action Optics — Chrome extension. Detects and visualizes which elements on a page an agent can actually operate.

7. Citing and correcting this site

Cite the specific ranking URL and its published or revised date, both of which appear at the top of every ranking. Scores change when evidence changes, and every change is logged.

Found an error? Corrections & Updates documents what to include and how it is handled. That page is the right target for a correction, not the contact form.