[JS] Implement Text2ImagePipeline

samples/js/image_generation/README.md

@@ -0,0 +1,84 @@
+# Text to Image JavaScript Generation Pipeline
+
+This example showcases inference of text-to-image diffusion models like Stable Diffusion 1.5, 2.1, FLUX, and LCM. The application doesn't have many configuration options to encourage the reader to explore and modify the source code. For example, change the device for inference to GPU. The sample features `Text2ImagePipeline` from `openvino-genai-node` and uses a text prompt as input source.
+
+Sample files:
+ - [`text2image.js`](./text2image.js) demonstrates basic usage of the text-to-image pipeline with a step callback and saves the result as a BMP file using `bmp-js`
+
+Users can change the sample code and play with the following generation parameters:
+
+- Change width or height of generated image
+- Generate multiple images per prompt (`num_images_per_prompt`)
+- Adjust a number of inference steps (`num_inference_steps`)
+- Play with [guidance scale](https://huggingface.co/spaces/stabilityai/stable-diffusion/discussions/9) (read [more details](https://arxiv.org/abs/2207.12598))
+- (SD 1.x, 2.x; SD3, SDXL) Add negative prompt when guidance scale > 1
+- (SDXL, SD3, FLUX) Specify other positive prompts like `prompt_2`
+- Add a per-step callback to monitor progress or stop generation early
+
+## Download and convert the model
+
+The `--upgrade-strategy eager` option is needed to ensure `optimum-intel` is upgraded to the latest version.
+
+It's not required to install [../../export-requirements.txt](../../export-requirements.txt) for deployment if the model has already been exported.
+
+```sh
+pip install --upgrade-strategy eager -r ../../export-requirements.txt
+```
+
+Then, run the export with Optimum CLI:
+
+```sh
+optimum-cli export openvino --model dreamlike-art/dreamlike-anime-1.0 --task stable-diffusion --weight-format fp16 dreamlike_anime_1_0_ov/FP16
+```
+
+## Run
+
+From the `samples/js` directory, install dependencies (if not already done):
+
+```bash
+npm install
+```
+
+If you use the master branch, you may need to [build openvino-genai-node from source](../../../src/js/README.md#build-bindings) first.
+
+Run the sample:
+
+```bash
+node image_generation/text2image.js dreamlike_anime_1_0_ov/FP16 "cyberpunk cityscape like Tokyo New York with tall buildings at dusk golden hour cinematic lighting"
+```
+
+The result is saved as `image.bmp` in the current directory.
+
+### Optional: change device
+
+The device is hardcoded to `CPU` in the sample. To use `GPU`, edit `text2image.js` and change:
+
+```js
+const device = "CPU"; // GPU can be used as well
+```
+
+### Examples
+
+Prompt: `cyberpunk cityscape like Tokyo New York with tall buildings at dusk golden hour cinematic lighting`
+
+   ![](./../../cpp/image_generation/512x512.bmp)
+
+Refer to the [Supported Models](https://openvinotoolkit.github.io/openvino.genai/docs/supported-models/#image-generation-models) for the list of supported models.
+
+## Run with a step callback
+
+The sample registers a callback that prints progress to stdout on each denoising step. The callback can also stop generation early by returning `true`:
+
+```js
+function callback(step, numSteps) {
+  process.stdout.write(`Step ${step + 1}/${numSteps}\r`);
+  return false; // return true to stop early
+}
+
+const imageTensor = await pipeline.generate(prompt, {
+  width: 512,
+  height: 512,
+  num_inference_steps: 20,
+  callback,
+});
+```

samples/js/image_generation/text2image.js

@@ -0,0 +1,80 @@
+// Copyright (C) 2023-2026 Intel Corporation
+// SPDX-License-Identifier: Apache-2.0
+
+import { writeFile } from "node:fs/promises";
+import { basename } from "node:path";
+import bmp from "bmp-js";
+import yargs from "yargs/yargs";
+import { hideBin } from "yargs/helpers";
+import { Text2ImagePipeline } from "openvino-genai-node";
+
+function toABGR(tensor) {
+  const [_, height, width, channels] = tensor.getShape();
+  if (channels !== 3) {
+    throw new Error(`Expected RGB image tensor, got ${channels} channels.`);
+  }
+
+  const rgb = tensor.data instanceof Uint8Array ? tensor.data : Uint8Array.from(tensor.data);
+  const rgba = Buffer.allocUnsafe(width * height * 4);
+
+  for (let src = 0, dst = 0; src < rgb.length; src += 3, dst += 4) {
+    rgba[dst] = 255;              // A
+    rgba[dst + 1] = rgb[src + 2]; // B
+    rgba[dst + 2] = rgb[src + 1]; // G
+    rgba[dst + 3] = rgb[src];     // R
+  }
+
+  return { height, width, rgba };
+}
+
+async function main() {
+  const argv = await yargs(hideBin(process.argv))
+    .scriptName(basename(process.argv[1]))
+    .command(
+      "$0 <model_dir> <prompt>",
+      "Run Text2Image pipeline and save generated image as BMP file",
+      (yargsBuilder) =>
+        yargsBuilder
+          .positional("model_dir", {
+            type: "string",
+            describe: "Path to the converted image generation model directory",
+            demandOption: true,
+          })
+          .positional("prompt", {
+            type: "string",
+            describe: "Prompt to generate images from",
+            demandOption: true,
+          })
+    )
+    .strict()
+    .help()
+    .parse();
+
+  const device = "CPU"; // GPU can be used as well
+  const pipeline = await Text2ImagePipeline(argv.model_dir, device);
+
+  function callback(step, numSteps) {
+    process.stdout.write(`Step ${step + 1}/${numSteps}\r`);
+    return false;
+  }
+
+  const imageTensor = await pipeline.generate(
+    argv.prompt,
+    {
+      width: 512,
+      height: 512,
+      num_inference_steps: 20,
+      num_images_per_prompt: 1,
+      callback,
+    },
+  );
+
+  const { height, width, rgba } = toABGR(imageTensor);
+  const bmpData = bmp.encode({ width, height, data: rgba });
+  await writeFile("image.bmp", bmpData.data);
+}
+
+main().catch((error) => {
+  console.error(error);
+  process.exitCode = 1;
+});

samples/js/package.json

@@ -4,9 +4,10 @@
   "license": "Apache-2.0",
   "type": "module",
   "devDependencies": {
+    "bmp-js": "^0.1.0",
     "node-wav": "^0.0.2",
-    "openvino-node": "^2026.1.0",
     "openvino-genai-node": "^2026.1.0",
+    "openvino-node": "^2026.1.0",
     "yargs": "^18.0.0",
     "zod": "^4.1.13"
   },

site/docs/bindings/node-js.md

@@ -9,7 +9,7 @@ description: Node.js bindings provide JavaScript/TypeScript API.
 OpenVINO GenAI provides Node.js bindings that enable you to use generative AI pipelines in JavaScript and TypeScript applications.
 
 :::warning API Coverage
-Node.js bindings currently provide a subset of the full OpenVINO GenAI API available in C++ and Python. The focus is on core text generation (`LLMPipeline`), vision language models (`VLMPipeline`), text embedding (`TextEmbeddingPipeline`), text reranking (`TextRerankPipeline`), speech recognition (`WhisperPipeline`), and speech generation (`Text2SpeechPipeline`) functionality.
+Node.js bindings currently provide a subset of the full OpenVINO GenAI API available in C++ and Python. The focus is on core text generation (`LLMPipeline`), vision language models (`VLMPipeline`), text embedding (`TextEmbeddingPipeline`), text reranking (`TextRerankPipeline`), speech recognition (`WhisperPipeline`), speech generation (`Text2SpeechPipeline`), and image generation (`Text2ImagePipeline`) functionality.
 :::
 
 ## Supported Pipelines and Features
@@ -36,6 +36,10 @@ Node.js bindings currently support:
 - `Text2SpeechPipeline`: Speech generation from text
   - Optional speaker embedding support
   - Batch generation
+- `Text2ImagePipeline`: Image generation from text prompts using diffusion models
+  - Configurable width, height, inference steps, and guidance scale
+  - Batch image generation via `num_images_per_prompt`
+  - Callback support for monitoring or stopping generation early
 - `Tokenizer`: Fast tokenization / detokenization and chat prompt formatting
   - Encode strings into token id and attention mask tensors
   - Decode token sequences

site/docs/use-cases/image-generation/_sections/_run_model/_text2image_js.mdx

@@ -0,0 +1,26 @@
+import CodeBlock from '@theme/CodeBlock';
+
+<CodeBlock language="javascript" showLineNumbers>
+{`import { writeFile } from 'node:fs/promises';
+import bmp from 'bmp-js';
+import { Text2ImagePipeline } from 'openvino-genai-node';
+
+const pipeline = await Text2ImagePipeline(modelPath, '${props.device || 'CPU'}');
+const image = await pipeline.generate(prompt);
+
+// Save the first generated image as BMP
+const [height, width] = [image.getShape()[1], image.getShape()[2]];
+const rgb = Uint8Array.from(image.data);
+const abgr = Buffer.allocUnsafe(width * height * 4);
+
+for (let src = 0, dst = 0; src < rgb.length; src += 3, dst += 4) {
+  abgr[dst] = 255;              // A
+  abgr[dst + 1] = rgb[src + 2]; // B
+  abgr[dst + 2] = rgb[src + 1]; // G
+  abgr[dst + 3] = rgb[src];     // R
+}
+
+const bmpData = bmp.encode({ width, height, data: abgr });
+await writeFile('image.bmp', bmpData.data);
+`}
+</CodeBlock>

site/docs/use-cases/image-generation/_sections/_run_model/index.mdx

@@ -1,5 +1,6 @@
 import Text2ImageCPP from './_text2image_cpp.mdx';
 import Text2ImagePython from './_text2image_python.mdx';
+import Text2ImageJS from './_text2image_js.mdx';
 
 import Image2ImageCPP from './_image2image_cpp.mdx';
 import Image2ImagePython from './_image2image_python.mdx';
@@ -44,6 +45,16 @@ OpenVINO GenAI supports the following diffusion model pipelines:
             </TabItem>
         </Tabs>
     </TabItemCpp>
+    <TabItemJS>
+        <Tabs groupId="device">
+            <TabItem label="CPU" value="cpu">
+                <Text2ImageJS device="CPU" />
+            </TabItem>
+            <TabItem label="GPU" value="gpu">
+                <Text2ImageJS device="GPU" />
+            </TabItem>
+        </Tabs>
+    </TabItemJS>
 </LanguageTabs>
 
 :::tip

site/docs/use-cases/image-generation/_sections/_usage_options/index.mdx

@@ -3,7 +3,7 @@ import GenerationConfigurationWorkflow from '@site/docs/use-cases/_shared/_gener
 ## Additional Usage Options
 
 :::tip
-Check out [Python](https://github.com/openvinotoolkit/openvino.genai/tree/master/samples/python/image_generation) and [C++](https://github.com/openvinotoolkit/openvino.genai/tree/master/samples/cpp/image_generation) image generation samples.
+Check out [Python](https://github.com/openvinotoolkit/openvino.genai/tree/master/samples/python/image_generation), [C++](https://github.com/openvinotoolkit/openvino.genai/tree/master/samples/cpp/image_generation), and [JavaScript](https://github.com/openvinotoolkit/openvino.genai/tree/master/samples/js/image_generation) image generation samples.
 :::
 
 ### Use Different Generation Parameters
@@ -60,6 +60,22 @@ You can adjust several parameters to control the image generation process, inclu
         }
         ```
     </TabItemCpp>
+    <TabItemJS>
+        ```javascript
+        import { Text2ImagePipeline } from 'openvino-genai-node';
+
+        const pipeline = await Text2ImagePipeline(modelPath, 'CPU');
+        const result = await pipeline.generate(prompt, {
+            // highlight-start
+            width: 512,
+            height: 512,
+            num_images_per_prompt: 1,
+            num_inference_steps: 30,
+            guidance_scale: 7.5,
+            // highlight-end
+        });
+        ```
+    </TabItemJS>
 </LanguageTabs>
 
 :::info Understanding Image Generation Parameters

site/src/pages/_sections/UseCasesSection/components/image-generation.tsx

@@ -1,10 +1,11 @@
 import Button from '@site/src/components/Button';
-import { LanguageTabs, TabItemCpp, TabItemPython } from '@site/src/components/LanguageTabs';
+import { LanguageTabs, TabItemCpp, TabItemPython, TabItemJS } from '@site/src/components/LanguageTabs';
 
 import UseCaseCard from './UseCaseCard';
 
 import CodeExampleCpp from '@site/docs/use-cases/image-generation/_sections/_run_model/_text2image_cpp.mdx';
 import CodeExamplePython from '@site/docs/use-cases/image-generation/_sections/_run_model/_text2image_python.mdx';
+import CodeExampleJS from '@site/docs/use-cases/image-generation/_sections/_run_model/_text2image_js.mdx';
 
 export const ImageGeneration = () => (
   <UseCaseCard>
@@ -30,6 +31,9 @@ export const ImageGeneration = () => (
         <TabItemCpp>
           <CodeExampleCpp />
         </TabItemCpp>
+        <TabItemJS>
+          <CodeExampleJS />
+        </TabItemJS>
       </LanguageTabs>
     </UseCaseCard.Code>
     <UseCaseCard.Actions>

src/js/eslint.config.cjs

@@ -90,6 +90,9 @@ module.exports = defineConfig([
             "begin_suppress_tokens",
             "suppress_tokens",
             "initial_prompt",
+            "num_inference_steps",
+            "num_images_per_prompt",
+            "guidance_scale",
           ],
         },
       ],

src/js/include/addon.hpp

@@ -12,12 +12,14 @@ struct AddonData {
     Napi::FunctionReference vlm_pipeline;
     Napi::FunctionReference text_rerank_pipeline;
     Napi::FunctionReference whisper_pipeline;
+    Napi::FunctionReference text2image_pipeline;
     Napi::FunctionReference tokenizer;
     Napi::FunctionReference perf_metrics;
     Napi::FunctionReference vlm_perf_metrics;
     Napi::FunctionReference whisper_perf_metrics;
     Napi::FunctionReference text2speech_pipeline;
     Napi::FunctionReference text2speech_perf_metrics;
+    Napi::FunctionReference text2image_perf_metrics;
     Napi::FunctionReference chat_history;
     Napi::FunctionReference reasoning_parser;
     Napi::FunctionReference deepseek_r1_reasoning_parser;