@sogni-ai/sogni-client 4.2.0-alpha.2 → 4.2.0-alpha.4

package/CHANGELOG.md

@@ -1,3 +1,17 @@
+# [4.2.0-alpha.4](https://github.com/Sogni-AI/sogni-client/compare/v4.2.0-alpha.3...v4.2.0-alpha.4) (2026-04-22)
+
+
+### Features
+
+* expand chat sogni media tools ([fb0d6c7](https://github.com/Sogni-AI/sogni-client/commit/fb0d6c77e4b73fbbe57d8b64b4a37bb610eb8c68))
+
+# [4.2.0-alpha.3](https://github.com/Sogni-AI/sogni-client/compare/v4.2.0-alpha.2...v4.2.0-alpha.3) (2026-04-20)
+
+
+### Bug Fixes
+
+* use thinkingComplexDefault in LLMModelInfo cost estimates ([162011f](https://github.com/Sogni-AI/sogni-client/commit/162011f698b56229629cea8d09fad1b3e313294f))
+
 # [4.2.0-alpha.2](https://github.com/Sogni-AI/sogni-client/compare/v4.2.0-alpha.1...v4.2.0-alpha.2) (2026-04-20)
 
 

package/README.md

@@ -18,7 +18,7 @@ Behind the scenes this SDK uses a WebSocket connection for communication between
 - 🎯 **Advanced Controls** - Fine-tune generation with samplers, schedulers, ControlNets, and more
 - 🤖 **LLM Text Generation** - Chat completions with streaming, multi-turn conversations, and thinking/reasoning mode via OpenAI-compatible API
 - 🔧 **LLM Tool Calling** - Define custom tools (functions) that the LLM can invoke during conversations for real-time data and actions
-- 🎨🎬🎵 **Sogni Platform Tools** - Generate images, videos, and music through natural language chat — the LLM detects media intent, enhances prompts, and calls Sogni's generation APIs automatically
+- 🎨🎬🎵 **Sogni Platform Tools** - Generate images, reference-guided image edits, videos, audio-driven videos, video transforms, and music through natural language chat
 - 👁️ **Vision Chat** - Multimodal image understanding with scene description, OCR, object detection, visual analysis, and multi-image comparison via Qwen3.6 VLM
 ## Migration notes
 ### v3.x.x to v4.x.x
@@ -691,13 +691,20 @@ const response = await sogni.chat.completions.create({
 
 ### Sogni Platform Tools — Generate Media via Chat
 
-Combine LLM intelligence with Sogni's media generation capabilities. The LLM detects when a user wants to create an image, video, or music, enhances the prompt, and calls Sogni's generation APIs — turning natural language into creative output:
+Combine LLM intelligence with Sogni's media generation capabilities. The SDK exposes six built-in public platform tools for chat completions:
 
-- **Image Generation** via tool call — "Create an image of a cyberpunk city at night"
-- **Video Generation** via tool call — "Generate a video of ocean waves at sunset"
-- **Music Generation** via tool call — "Compose a jazz song about the rain"
+- **`sogni_generate_image`** — text-to-image generation
+- **`sogni_edit_image`** — reference-guided image editing using `source_image_url` and `reference_image_urls`
+- **`sogni_generate_video`** — text-to-video and image-to-video generation
+- **`sogni_sound_to_video`** — audio-driven video generation using `reference_audio_url`
+- **`sogni_video_to_video`** — video transformation / motion transfer using `reference_video_url`
+- **`sogni_generate_music`** — music generation with optional lyrics and advanced controls
 
-See the `workflow_text_chat_sogni_tools.mjs` example for a complete implementation that wires LLM tool calling to Sogni's image, video, and audio generation APIs.
+Use `SogniTools.all` to expose the full tool surface, then execute tool calls with `sogni.chat.tools.execute()` / `executeAll()` or `autoExecuteTools: true` for non-streaming flows.
+
+Media-conditioned workflows use explicit asset URLs or data URIs, including `source_image_url`, `reference_image_url`, `reference_audio_url`, `reference_audio_identity_url`, and `reference_video_url`.
+
+The `workflow_text_chat_sogni_tools.mjs` example demonstrates the core text-to-image, text-to-video, and text-to-music composition flows. Dedicated workflow examples like `workflow_image_edit.mjs`, `workflow_sound_to_video.mjs`, and `workflow_video_to_video.mjs` cover the asset-backed workflows directly.
 
 ## Code Examples
 
@@ -719,7 +726,7 @@ The [examples](https://github.com/Sogni-AI/sogni-client/tree/main/examples) dire
 - **`workflow_text_chat_multi_turn.mjs`** - Multi-turn conversation with history, in-chat commands, and session stats
 - **`workflow_text_chat_vision.mjs`** - Vision chat with multimodal image understanding (scene description, OCR, object detection, visual analysis, multi-image comparison)
 - **`workflow_text_chat_tool_calling.mjs`** - LLM tool calling with built-in tools (weather, time, unit conversion, math)
-- **`workflow_text_chat_sogni_tools.mjs`** - Generate images, videos, and music through natural language via LLM tool calling
+- **`workflow_text_chat_sogni_tools.mjs`** - Core image/video/music generation through natural language via LLM tool calling
 
 ### Basic Examples
 - **`promise_based.mjs`** - Image generation using promises/async-await
@@ -736,7 +743,7 @@ The workflow examples showcase a few powerful open-source frontier models suppor
 | `qwen_image_edit_2511_fp8_lightning` | **Qwen Image Edit Lightning** - Fast 4-step editing | Rapid reference-based image generation |
 | `qwen_image_edit_2511_fp8` | **Qwen Image Edit** - High quality 20-step editing | Professional image editing with context awareness |
 | `wan_v2.2-14b-fp8_t2v_lightx2v` | **Wan 2.2 T2V** - Text-to-video | Generate videos from text prompts |
-| `qwen3.6-35b-a3b-gguf-iq4xs` | **Qwen3.6 35B VLM** - LLM chat, tool calling & vision | Latest model with 128K context target, reasoning, tool calling, and multimodal image understanding |
+| `qwen3.6-35b-a3b-gguf-iq4xs` | **Qwen3.6 35B VLM** - LLM chat, tool calling & vision | Latest model with 262,144 native context length, reasoning, tool calling, and multimodal image understanding |
 
 All workflow examples include:
 - Interactive model and parameter selection
@@ -777,7 +784,7 @@ When helping users generate images, videos, or use LLM features with Sogni:
 2. **Video generation**: Use `type: 'video'` with `network: 'fast'` (required)
 3. **Audio generation**: Use `type: 'audio'` with ACE-Step 1.5 models
 4. **LLM text chat**: Use `sogni.projects.chatCompletion()` for text generation with streaming and tool calling
-5. **Sogni Platform Tools**: Combine LLM tool calling with Sogni media generation to create images, videos, and music from natural language
+5. **Sogni Platform Tools**: Combine LLM tool calling with Sogni media generation to create images, image edits, videos, audio-driven videos, video transforms, and music from natural language
 6. **Vision chat**: Use `qwen3.6-35b-a3b-gguf-iq4xs` VLM for multimodal image understanding with `image_url` content type
 7. **WAN 2.2 vs LTX-2.3**: These model families have different FPS behaviors - see `llms-full.txt` for details
 

package/dist/Chat/ChatTools.d.ts

@@ -1,65 +1,21 @@
 import type ProjectsApi from '../Projects';
 import { ToolCall, ToolExecutionOptions, ToolExecutionProgress, ToolExecutionResult } from './types';
-/**
- * API for executing Sogni platform tool calls (image, video, music generation).
- *
- * Accessed via `sogni.chat.tools`. Provides methods to execute tool calls returned
- * by the LLM, mapping them to `sogni.projects.create()` calls automatically.
- *
- * @example
- * ```typescript
- * // Execute a single tool call
- * const result = await sogni.chat.tools.execute(toolCall, {
- *   tokenType: 'sogni',
- *   onProgress: (p) => console.log(`${p.status}: ${p.percent}%`),
- * });
- *
- * // Execute all tool calls from a completion
- * const results = await sogni.chat.tools.executeAll(result.tool_calls, {
- *   onToolCall: async (tc) => myCustomHandler(tc), // for non-Sogni tools
- * });
- * ```
- */
 declare class ChatToolsApi {
     private projects;
     constructor(projects: ProjectsApi);
-    /**
-     * Execute a single Sogni platform tool call.
-     *
-     * Maps tool call arguments to `sogni.projects.create()`, waits for the media
-     * generation to complete, and returns the result URLs.
-     *
-     * @throws Error if the tool call is not a Sogni tool (use `isSogniToolCall()` to check first)
-     */
     execute(toolCall: ToolCall, options?: ToolExecutionOptions): Promise<ToolExecutionResult>;
-    /**
-     * Execute multiple tool calls from a single LLM response.
-     *
-     * Sogni tool calls (prefixed with `sogni_`) are executed automatically via
-     * `projects.create()`. Non-Sogni tool calls are delegated to the `onToolCall`
-     * callback if provided, or returned as errors.
-     *
-     * @param toolCalls - Array of tool calls from `result.tool_calls`
-     * @param options - Execution options plus optional handler for non-Sogni tools
-     */
     executeAll(toolCalls: ToolCall[], options?: ToolExecutionOptions & {
-        /** Handler for non-Sogni tool calls. Must return the tool result content string. */
         onToolCall?: (toolCall: ToolCall) => Promise<string>;
-        /** Per-tool progress callback (wraps the per-tool onProgress with tool identity). */
         onToolProgress?: (toolCall: ToolCall, progress: ToolExecutionProgress) => void;
     }): Promise<ToolExecutionResult[]>;
-    /**
-     * Get the default model for a media type. For video, prefers LTX-2.3 t2v models.
-     * Falls back to the model with the most available workers.
-     */
-    private getDefaultModel;
-    /**
-     * Create a project, wait for completion with timeout, track per-job progress,
-     * and clean up on failure or timeout.
-     */
+    private getAvailableModels;
+    private selectModel;
     private executeProject;
     private executeImageGeneration;
+    private executeImageEdit;
     private executeVideoGeneration;
+    private executeSoundToVideo;
+    private executeVideoToVideo;
     private executeMusicGeneration;
     private makeErrorResult;
 }