TanStack Docs Language Inventory

This recorded inventory was generated from local Markdown and MDX files in:

Sibling repositories under ../*/docs
The sibling ../tanstack.com repository

The scan parses fenced code blocks statefully, so closing fences are not counted as plaintext blocks. Hidden directories and this highlight package are ignored.

Scanned files in this snapshot: 2940. Counts depend on the sibling checkouts and are not a current inventory of every TanStack repository. Run pnpm run scan:languages from a checkout with the source repositories available to produce a new inventory.

Generated real-code fixtures:

shell

pnpm run extract:fixtures

The current fixture set stores up to twenty sampled code fences per normalized language target in test/generated/tanstack-doc-fixtures.json.

Raw Fence Tags

Count	Tag
8173	ts
4078	tsx
904	js
877	typescript
561	bash
226	plaintext
211	json
136	sh
107	vue
96	jsx
86	svelte
51	angular-ts
43	text
38	txt
33	mermaid
33	shell
28	html
26	js-vue
21	yaml
15	toml
14	javascript
11	css
11	markdown
11	md
11	sql
9	angular-html
8	http
6	env
4	dockerfile
4	nginx
4	python
3	diff
3	xml
2	apache
2	dotenv
2	ejs
1	-->
1	jsonc
1	scheme

Normalized Support Targets

--> is malformed markdown and is normalized to plaintext.

Target	Aliases
tsx	tsx
ts	ts, typescript, angular-ts
jsx	jsx
js	js, javascript, js-vue
shell	bash, sh, shell
json	json, jsonc, jsonl
plaintext	plaintext, text, txt, -->
vue	vue
svelte	svelte
html	html, angular-html, xml
markdown	md, markdown
yaml	yaml
css	css
diff	diff
mermaid	mermaid
toml	toml
sql	sql
http	http
env	env, dotenv
dockerfile	dockerfile
nginx	nginx
python	python
apache	apache
ejs	ejs
scheme	scheme

Contract

The highlighter output should be one HTML tree with stable semantic classes. Theme switching should happen in CSS, not by duplicating HTML, embedding inline colors, or re-tokenizing for light and dark themes.