@push.rocks/smartai 0.7.0 → 0.7.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -3,7 +3,7 @@
3
3
  */
4
4
  export const commitinfo = {
5
5
  name: '@push.rocks/smartai',
6
- version: '0.7.0',
6
+ version: '0.7.1',
7
7
  description: 'SmartAi is a versatile TypeScript library designed to facilitate integration and interaction with various AI models, offering functionalities for chat, audio generation, document processing, and vision tasks.'
8
8
  };
9
9
  //# sourceMappingURL=data:application/json;base64,eyJ2ZXJzaW9uIjozLCJmaWxlIjoiMDBfY29tbWl0aW5mb19kYXRhLmpzIiwic291cmNlUm9vdCI6IiIsInNvdXJjZXMiOlsiLi4vdHMvMDBfY29tbWl0aW5mb19kYXRhLnRzIl0sIm5hbWVzIjpbXSwibWFwcGluZ3MiOiJBQUFBOztHQUVHO0FBQ0gsTUFBTSxDQUFDLE1BQU0sVUFBVSxHQUFHO0lBQ3hCLElBQUksRUFBRSxxQkFBcUI7SUFDM0IsT0FBTyxFQUFFLE9BQU87SUFDaEIsV0FBVyxFQUFFLGtOQUFrTjtDQUNoTyxDQUFBIn0=
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@push.rocks/smartai",
3
- "version": "0.7.0",
3
+ "version": "0.7.1",
4
4
  "private": false,
5
5
  "description": "SmartAi is a versatile TypeScript library designed to facilitate integration and interaction with various AI models, offering functionalities for chat, audio generation, document processing, and vision tasks.",
6
6
  "main": "dist_ts/index.js",
package/readme.md CHANGED
@@ -45,15 +45,15 @@ const response = await ai.openaiProvider.chat({
45
45
 
46
46
  Choose the right provider for your use case:
47
47
 
48
- | Provider | Chat | Streaming | TTS | Vision | Documents | Research | Highlights |
49
- |----------|:----:|:---------:|:---:|:------:|:---------:|:--------:|------------|
50
- | **OpenAI** | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | • GPT-4, DALL-E 3<br>• Industry standard<br>• Deep research API |
51
- | **Anthropic** | ✅ | ✅ | ❌ | ✅ | ✅ | ✅ | • Claude 3 Opus<br>• Superior reasoning<br>• Web search API |
52
- | **Ollama** | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | • 100% local<br>• Privacy-first<br>• No API costs |
53
- | **XAI** | ✅ | ✅ | ❌ | ❌ | ✅ | ❌ | • Grok models<br>• Real-time data<br>• Uncensored |
54
- | **Perplexity** | ✅ | ✅ | ❌ | ❌ | ❌ | ✅ | • Web-aware<br>• Research-focused<br>• Sonar Pro models |
55
- | **Groq** | ✅ | ✅ | ❌ | ❌ | ❌ | ❌ | • 10x faster<br>• LPU inference<br>• Low latency |
56
- | **Exo** | ✅ | ✅ | ❌ | ❌ | ❌ | ❌ | • Distributed<br>• P2P compute<br>• Decentralized |
48
+ | Provider | Chat | Streaming | TTS | Vision | Documents | Research | Images | Highlights |
49
+ |----------|:----:|:---------:|:---:|:------:|:---------:|:--------:|:------:|------------|
50
+ | **OpenAI** | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | gpt-image-1<br>• DALL-E 3<br>• Deep research API |
51
+ | **Anthropic** | ✅ | ✅ | ❌ | ✅ | ✅ | ✅ | ❌ | • Claude 3 Opus<br>• Superior reasoning<br>• Web search API |
52
+ | **Ollama** | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ | • 100% local<br>• Privacy-first<br>• No API costs |
53
+ | **XAI** | ✅ | ✅ | ❌ | ❌ | ✅ | ❌ | ❌ | • Grok models<br>• Real-time data<br>• Uncensored |
54
+ | **Perplexity** | ✅ | ✅ | ❌ | ❌ | ❌ | ✅ | ❌ | • Web-aware<br>• Research-focused<br>• Sonar Pro models |
55
+ | **Groq** | ✅ | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ | • 10x faster<br>• LPU inference<br>• Low latency |
56
+ | **Exo** | ✅ | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ | • Distributed<br>• P2P compute<br>• Decentralized |
57
57
 
58
58
  ## 🎮 Core Features
59
59
 
@@ -216,6 +216,87 @@ const perplexityResearch = await ai.perplexityProvider.research({
216
216
  - **Anthropic**: Web Search API with domain filtering
217
217
  - **Perplexity**: Sonar and Sonar Pro models with built-in citations
218
218
 
219
+ ### 🎨 Image Generation & Editing
220
+
221
+ Generate and edit images with OpenAI's cutting-edge models:
222
+
223
+ ```typescript
224
+ // Basic image generation with gpt-image-1
225
+ const image = await ai.openaiProvider.imageGenerate({
226
+ prompt: 'A futuristic robot assistant in a modern office, digital art',
227
+ model: 'gpt-image-1',
228
+ quality: 'high',
229
+ size: '1024x1024'
230
+ });
231
+
232
+ // Save the generated image
233
+ const imageBuffer = Buffer.from(image.images[0].b64_json!, 'base64');
234
+ fs.writeFileSync('robot.png', imageBuffer);
235
+
236
+ // Advanced: Transparent background with custom format
237
+ const logo = await ai.openaiProvider.imageGenerate({
238
+ prompt: 'Minimalist mountain peak logo, geometric design',
239
+ model: 'gpt-image-1',
240
+ quality: 'high',
241
+ size: '1024x1024',
242
+ background: 'transparent',
243
+ outputFormat: 'png'
244
+ });
245
+
246
+ // WebP with compression for web use
247
+ const webImage = await ai.openaiProvider.imageGenerate({
248
+ prompt: 'Product showcase: sleek smartphone on marble surface',
249
+ model: 'gpt-image-1',
250
+ quality: 'high',
251
+ size: '1536x1024',
252
+ outputFormat: 'webp',
253
+ outputCompression: 85
254
+ });
255
+
256
+ // Superior text rendering (gpt-image-1's strength)
257
+ const signage = await ai.openaiProvider.imageGenerate({
258
+ prompt: 'Vintage cafe sign saying "COFFEE & CODE" in hand-lettered typography',
259
+ model: 'gpt-image-1',
260
+ quality: 'high',
261
+ size: '1024x1024'
262
+ });
263
+
264
+ // Generate multiple variations at once
265
+ const variations = await ai.openaiProvider.imageGenerate({
266
+ prompt: 'Abstract geometric pattern, colorful minimalist art',
267
+ model: 'gpt-image-1',
268
+ n: 3,
269
+ quality: 'medium',
270
+ size: '1024x1024'
271
+ });
272
+
273
+ // Edit an existing image
274
+ const editedImage = await ai.openaiProvider.imageEdit({
275
+ image: originalImageBuffer,
276
+ prompt: 'Add sunglasses and change the background to a beach sunset',
277
+ model: 'gpt-image-1',
278
+ quality: 'high'
279
+ });
280
+ ```
281
+
282
+ **Image Generation Options:**
283
+ - `model`: 'gpt-image-1' | 'dall-e-3' | 'dall-e-2'
284
+ - `quality`: 'low' | 'medium' | 'high' | 'auto'
285
+ - `size`: Multiple aspect ratios up to 4096×4096
286
+ - `background`: 'transparent' | 'opaque' | 'auto'
287
+ - `outputFormat`: 'png' | 'jpeg' | 'webp'
288
+ - `outputCompression`: 0-100 for webp/jpeg
289
+ - `moderation`: 'low' | 'auto'
290
+ - `n`: Number of images (1-10)
291
+
292
+ **gpt-image-1 Advantages:**
293
+ - Superior text rendering in images
294
+ - Up to 4096×4096 resolution
295
+ - Transparent background support
296
+ - Advanced output formats (WebP with compression)
297
+ - Better prompt understanding
298
+ - Streaming support for progressive rendering
299
+
219
300
  ### 🔄 Persistent Conversations
220
301
 
221
302
  Maintain context across interactions:
@@ -501,8 +582,8 @@ export PERPLEXITY_API_KEY=pplx-...
501
582
  ## 📈 Roadmap
502
583
 
503
584
  - [x] Research & Web Search API
585
+ - [x] Image generation support (gpt-image-1, DALL-E 3, DALL-E 2)
504
586
  - [ ] Streaming function calls
505
- - [ ] Image generation support
506
587
  - [ ] Voice input processing
507
588
  - [ ] Fine-tuning integration
508
589
  - [ ] Embedding support
@@ -3,6 +3,6 @@
3
3
  */
4
4
  export const commitinfo = {
5
5
  name: '@push.rocks/smartai',
6
- version: '0.7.0',
6
+ version: '0.7.1',
7
7
  description: 'SmartAi is a versatile TypeScript library designed to facilitate integration and interaction with various AI models, offering functionalities for chat, audio generation, document processing, and vision tasks.'
8
8
  }