v4.1.0

- Update packages
- Update provider list
- Fix type definitions
- Fix some minor issues
- Add AI Agents Guides

Author: Dong Nguyen

.aiignore

@@ -0,0 +1,15 @@
+node_modules
+coverage
+coverage.lcov
+
+package-lock.json
+pnpm-lock.yaml
+bun.lock
+
+.env
+
+dist
+storage
+
+# AI Session Files (Private Context)
+.sessions

.gitignore

@@ -16,3 +16,4 @@ yarn.lock
 pnpm-lock.yaml
 package-lock.json
 deno.lock
+bun.lock

AGENTS.md

@@ -0,0 +1,114 @@
+# AI Agent Instructions
+
+Coding guidelines for AI agents working in this project.
+
+## Philosophy
+
+- Minimalism. Simple is better. KISS (Keep It Simple, Stupid).
+- Clean code, easy to read, easy to delete.
+- Functional Programming — pure functions, immutability, no side effects.
+- MVP mindset — deliver the smallest thing that works, then iterate.
+
+## Security Rules (CRITICAL — no exceptions)
+
+- NEVER output or request .env and example.env file contents
+- NEVER hardcode API credentials, secret tokens, private keys or passwords in source code
+- NEVER send sensitive data to external AI services
+- Follow `.aiignore` and `.gitignore` for excluded files — do not read or reference them
+- When asking for help, sanitize data (replace real IDs, emails, tokens with placeholders)
+- Do not log sensitive information
+
+## Coding Standards (Strict)
+- Language: JavaScript (ESM syntax). No TypeScript.
+- Style: No semicolons, single quotes, 2-space indentation.
+- Respect `eslint.config.js` — do not suggest rule changes
+- Patterns:
+  - Functional Programming only. No Classes or OOP.
+  - Arrow functions are preferred.
+  - Maximum 3 parameters per function. Use objects for more.
+- Naming: camelCase for variables/functions, SNAKE_CASE for constants.
+- Documentation:
+  - Add JSDocs before all functions and exported variables.
+  - Language: Use American English for all comments and JSDocs.
+  - Constraint: NEVER use Vietnamese or other languages in the source code.
+
+### Error Handling
+
+- Handle errors explicitly — never swallow silently
+- Use try/catch with proper logging
+- Return null or throw meaningful errors
+
+```javascript
+export const send = async (params) => {
+  try {
+    const response = await ai.ask(params)
+    logger.info(`send() -> success: ${response.id}`)
+    return response
+  } catch (err) {
+    logger.error(`send() -> failed: ${err.message}`)
+    console.error(err)
+    return null
+  }
+}
+```
+
+## Testing Standards
+
+- Write tests for critical business logic, all error cases
+- Use simple test runners (node:test, bun:test, vitest)
+- No complex mocking frameworks unless necessary
+- Tests live alongside source: `[module].test.js` next to `[module].js`
+
+## Dependency Rules
+
+- Prefer built-in APIs over external packages
+- Before adding dependency, explain:
+  - Why it is needed
+  - Alternatives considered
+  - Bundle size impact
+- Never add dependency for trivial utilities
+- Avoid packages with large dependency trees
+
+## Architecture Rules
+
+- Do NOT change existing project architecture without explicit approval
+- Do NOT move or rename core modules unless requested
+- Respect module boundaries
+- Avoid cross-module coupling
+- New modules must follow existing folder structure
+
+## When Making Changes
+
+1. Read existing patterns first
+2. Follow current coding style strictly
+3. Keep dependencies minimal
+4. Handle errors explicitly
+5. Add JSDoc comments for new functions
+6. Run `npm run lint` before committing
+7. Do NOT refactor unrelated code
+8. Do NOT modify working code outside task scope
+9. Prefer minimal diff changes
+10. Preserve existing behavior unless explicitly requested
+
+## When in Doubt
+
+- Ask for clarification before generating code
+- State your assumption explicitly if proceeding without confirmation
+- Prefer doing less and asking over doing more and guessing
+
+## Git Workflow
+
+- Work only inside the current branch
+- Do NOT create or delete branches
+- Do NOT rewrite git history
+- Do NOT modify commit messages
+- Changes must correspond to the current issue
+
+## Agent References
+
+Reference these URLs when working on related topics:
+
+- Bun: https://bun.sh/llms-full.txt
+- oEmbed: https://oembed.com
+
+---

README.md

@@ -9,17 +9,21 @@ Extract oEmbed content from given URL.
 
 ## Demo
 
-- [Give it a try!](https://extractus-demo.vercel.app/oembed)
+- [Give it a try!](https://extractus.pwshub.com/oembed)
 
 ## Install & Usage
 
-### Node.js
+### Bun & Node.js
 
 ```bash
+# bun
+bun add @extractus/oembed-extractor
+
+# npm
 npm i @extractus/oembed-extractor
 
 # pnpm
-pnpm i @extractus/oembed-extractor
+pnpm install @extractus/oembed-extractor
 
 # yarn
 yarn add @extractus/oembed-extractor
@@ -38,12 +42,6 @@ console.log(result)
 import { extract } from 'npm:@extractus/oembed-extractor'
 ```
 
-### Browser
-
-```ts
-import { extract } from "https://esm.sh/@extractus/oembed-extractor@latest"
-```
-
 ## APIs
 
 ### `.extract()`

eval.js

@@ -3,6 +3,11 @@
 
 import { extract } from './src/main.js'
 
+/**
+ * Extract and log oEmbed data from a URL.
+ *
+ * @param {string} url - URL to extract oEmbed from
+ */
 const run = async (url) => {
   try {
     console.time('extract-oembed')
@@ -14,6 +19,12 @@ const run = async (url) => {
   }
 }
 
+/**
+ * Parse CLI arguments and run extraction.
+ *
+ * @param {Array} argv - Process argument array
+ * @returns {Promise|string} Extraction promise or info message
+ */
 const init = (argv) => {
   if (argv.length === 3) {
     const url = argv[2]