> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bodyguard.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# AGENTS

# Mintlify documentation

## Working relationship

* You can push back on ideas-this can lead to better documentation. Cite sources and explain your reasoning when you do so
* ALWAYS ask for clarification rather than making assumptions
* NEVER lie, guess, or make up anything

## Project context

* Mintlify docs site for **Bodyguard** (contextual AI content moderation)
* Format: MDX files with YAML frontmatter
* Config: `docs.json` drives navigation, theme, languages, redirects
* Bilingual site: English under `en/`, French under `fr/`
* API reference "Reference" group is auto-generated from `openapi/analyze.openapi.yaml`
* Components: Mintlify built-ins (`Note`, `Warning`, `Card`, `CardGroup`, `CodeGroup`)

## Repository layout

```
AGENTS.md           — this file
docs.json           — navigation, theme, languages, redirects
.mintignore         — drafts/, *.draft.mdx excluded
en/                 — English content
  documentation/    — product docs (Getting started, Taxonomy)
  api-reference/    — hand-written API guides (introduction, reseller)
fr/                 — French mirror of en/
openapi/
  analyze.openapi.yaml  — drives API "Reference" group
images/             — logo-light.svg, logo-dark.svg, favicon.svg
```

## Product & API facts

* Product: Bodyguard — classifies text, images, and usernames for harmful content
* API base: `https://api.bodyguard.ai`
* Auth header: `Authorization: ApiKey <your-api-key>` (scheme name is case-sensitive)
* Endpoints: `/analyze/v1/text`, `/analyze/v1/image`, `/analyze/v1/username`
* Text taxonomy has two layers:
  * Moderation signals: `type`, `severity`, `directedAt`
  * Classifications: **General** (everyone), **Additional** (Advanced plan, monitoring-only), **Custom** (per contract)
* Image taxonomy is flat — each classification stands on its own
* Reseller API keys unlock an `analysisSettings` object in request bodies (`industry`, `defaultLanguage`, `imageVlmClassifications`, `imageVlmSettings`, `imageOcrSettings`)
* Error envelope: `{ "error": { "code": "...", "details": { "<camelCaseOfCode>": { ... } } } }`

## Bilingual content

* Prefer mirroring every `en/<path>.mdx` at `fr/<path>.mdx` with identical relative path
* When adding or renaming a page, update **both** language blocks in `docs.json` navigation
* If the French mirror is pending, flag the missing mirror to the user rather than silently skipping it
* Translate frontmatter (`title`, `description`) and body; keep technical values (classification names, field names, code) unchanged

## Navigation registration (`docs.json`)

* New pages are invisible until added to `navigation.languages[].tabs[].groups[].pages[]`
* Page path format: `en/path/to/page` — no leading slash, no `.mdx` extension
* `openapi` groups auto-populate from the YAML file — do not list those endpoints manually
* Refer to the [docs.json schema](https://mintlify.com/docs.json) when editing navigation
* Respect existing tab names per language: English uses "Documentation" / "API Reference", French uses "Documentation" / "Référence API"

## Components

* You can use **any Mintlify built-in component**. Full catalog: [https://mintlify.com/docs/components](https://mintlify.com/docs/components)
* Prefer existing components over custom markup. If unsure a component exists, check the Mintlify docs before inventing one
* Icon library is Lucide (set in `docs.json`) — icon names must match Lucide

### Components already in use in this repo

Use these first for consistency; reach for others from the Mintlify catalog when they fit better:

* Callouts: `<Note>`, `<Warning>`
* Indexes and catalogs: `<CardGroup cols={2|3}>` containing `<Card title="..." icon="..." href="/en/...">`
* Multi-language code examples: `<CodeGroup>` with language-tagged fences (` ```bash cURL `, ` ```javascript JavaScript `, ` ```python Python `, ` ```go Go `)
* Card meta: `tag=` attribute — values seen in this repo: `"Additional"`, `"Removed by default"`, `"Never removed"`, `"Critical"`, `"Mutually exclusive"`

### Other useful Mintlify components (non-exhaustive)

`<Tip>`, `<Info>`, `<Check>`, `<Steps>` / `<Step>`, `<Accordion>` / `<AccordionGroup>`, `<Tabs>` / `<Tab>`, `<Expandable>`, `<Frame>`, `<Columns>`, `<Tooltip>`, `<Update>`, `<ParamField>`, `<ResponseField>`, `<RequestExample>`, `<ResponseExample>`. See the Mintlify components docs for the full list, props, and examples.

## Taxonomy page conventions

* Each classification rendered as a `<Card>` inside a grouped `<CardGroup>`
* Card body structure:
  1. `` `TECHNICAL_NAME` · Type `` (type only for text taxonomy — image taxonomy is untyped)
  2. One-sentence description
  3. Italic example in quotes (`*"..."*`)
* Group classifications with `###` subheadings. Existing groupings:
  * Text: Harmful content, Explicit content, Spam & low quality, Personal information (PII), Advertising, Positive signals, Brand & community, Context & topics
  * Image: Sexual content, Violence & harm, Hate & harassment, Substances & regulated activity, People & identity, Codes & information, Other

## Style anchors

* API reference pages: match `en/api-reference/introduction.mdx` (auth, versioning, rate limits, errors, Next steps CardGroup)
* Reseller / advanced API pages: match `en/api-reference/reseller.mdx`
* Taxonomy pages: match `en/documentation/taxonomy/text-classifications.mdx`

## Frontmatter requirements for pages

* `title`: Clear, descriptive page title
* `description`: Concise summary for SEO/navigation

## Writing standards

* Second-person voice ("you")
* Prerequisites at start of procedural content
* Test all code examples before publishing
* API code examples must stay consistent with `openapi/analyze.openapi.yaml` (endpoint path, field names, example values)
* Match style and formatting of existing pages — see Style anchors above
* Include both basic and advanced use cases
* Language tags on all code blocks
* Alt text on all images
* Relative paths for internal links

## Content strategy

* Document just enough for user success - not too much, not too little
* Prioritize accuracy and usability
* Make content evergreen when possible
* Search for existing content before adding anything new. Avoid duplication unless it is done for a strategic reason
* Check existing patterns for consistency
* Start by making the smallest reasonable changes

## Local preview

```
npm i -g mint
mint dev
```

Run at repo root (where `docs.json` lives). Preview at `http://localhost:3000`.

## Checklist before creating new content

1. Grep `en/` for existing similar page — avoid duplication
2. Confirm target tab and group in `docs.json` with the user if unclear
3. Draft `en/<path>.mdx` following the matching Style anchor
4. Mirror to `fr/<path>.mdx` (or flag the pending mirror)
5. Register the new path in **both** language blocks of `docs.json`
6. Run `mint dev` to verify rendering and links

## Do not

* Skip frontmatter on any MDX file
* Use absolute URLs for internal links
* Include untested code examples
* List OpenAPI-generated endpoints manually in `docs.json`
* Add a page without registering it in `docs.json` navigation
* Make assumptions - always ask for clarification