How to add custom features to the ONLYOFFICE AI plugin

The ONLYOFFICE AI plugin has introduced a new, developer-friendly architecture. If you want to extend the plugin capabilities, for example, to add a specific AI function, you no longer need to edit a massive, monolithic file.

Instead, the plugin now features a dedicated workspace located in the .dev folder. This guide will show you exactly how to use this workspace to add a new custom function.

We will build the Describe image feature for the Document Editor as our example.

The development workflow

The core concept is simple: You work in .dev, and a script builds the production code.

• .dev/helpers/: This is your sandbox. This is where you create new files.
• helpers.js: This is the production file the plugin actually reads. Do not edit this directly. It is auto-generated.

Step 1: Choose your editor scope

Inside .dev/helpers/, you will find three folders corresponding to the ONLYOFFICE editors:

• word/ (Document Editor)
• cell/ (Spreadsheet Editor)
• slide/ (Presentation Editor)

Since our describe image function is for text documents, we will work in .dev/helpers/word/.

Step 2: Create your own function

Create a new file named describe-image.js inside .dev/helpers/word/.

The plugin uses a specific class called RegisteredFunction to define capabilities. This structure tells the AI what the function does and how to call it.

Tip: Ensure your syntax is clean (watch for missing closing brackets).

The code:

(function () {
  let func = new RegisteredFunction({
    name: "describeImage",
    description:
      "Allows users to select an image and generate a meaningful title, description, caption, or alt text for it using AI.",
    // Define parameters so the AI knows what to ask for
    parameters: {
      type: "object",
      properties: {
        prompt: {
          type: "string",
          description:
            "instruction for the AI (e.g., 'Add a short title for this chart.')",
        },
      },
      required: ["prompt"],
    },
    // Provide examples to train the AI on usage
    examples: [
      {
        prompt: "Add a short title for this chart.",
        arguments: { prompt: "Add a short title for this chart." },
      },
      {
        prompt: "Write me a 1–2 sentence description of this photo.",
        arguments: {
          prompt: "Write me a 1–2 sentence description of this photo.",
        },
      },
      {
        prompt: "Generate a descriptive caption for this organizational chart.",
        arguments: {
          prompt:
            "Generate a descriptive caption for this organizational chart.",
        },
      },
      {
        prompt: "Provide accessibility-friendly alt text for this infographic.",
        arguments: {
          prompt:
            "Provide accessibility-friendly alt text for this infographic.",
        },
      },
    ],
  });

   // The actual logic executed inside the editor
  func.call = async function (params) {
    let prompt = params.prompt;

    async function insertMessage(message) {
      Asc.scope._message = String(message || "");
      // 3. Insert the result into the document
      await Asc.Editor.callCommand(function () {
        const msg = Asc.scope._message || "";
        const doc = Api.GetDocument();
        const selected =
          (doc.GetSelectedDrawings && doc.GetSelectedDrawings()) || [];
        if (selected.length > 0) {
          for (let i = 0; i < selected.length; i++) {
            const drawing = selected[i];
            const para = Api.CreateParagraph();
            para.AddText(msg);
            drawing.InsertParagraph(para, "after", true);
          }
        } else {
          const para = Api.CreateParagraph();
          para.AddText(msg);
          let range = doc.GetCurrentParagraph();
          range.InsertParagraph(para, "after", true);
        }
        Asc.scope._message = "";
      }, true);
    }

    try {
    // 1. Get the selected image
      let imageData = await new Promise((resolve) => {
        window.Asc.plugin.executeMethod(
          "GetImageDataFromSelection",
          [],
          function (result) {
            resolve(result);
          }
        );
      });

      if (!imageData || !imageData.src) {
        await insertMessage("Please select a valid image first.");
        return;
      }

      const whiteRectangleBase64 =
        "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mP8/5+hHgAHggJ/PchI7wAAAABJRU5ErkJggg==";
      if (imageData.src === whiteRectangleBase64) {
        await insertMessage("Please select a valid image first.");
        return;
      }

      let argPrompt = prompt + " (for the selected image)";
      // 2. Send image + prompt to the AI engine
      let requestEngine = AI.Request.create(AI.ActionType.Chat);
      if (!requestEngine) {
        await insertMessage("AI request engine not available.");
        return;
      }

      const allowVision = /(vision|gemini|gpt-4o|gpt-4v|gpt-4-vision)/i;
      if (!allowVision.test(requestEngine.modelUI.name)) {
        await insertMessage(
          "⚠ This model may not support images. Please choose a vision-capable model (e.g. GPT-4V, Gemini, etc.)."
        );
        return;
      }

      await Asc.Editor.callMethod("StartAction", [
        "Block",
        "AI (" + requestEngine.modelUI.name + ")",
      ]);
      await Asc.Editor.callMethod("StartAction", ["GroupActions"]);

      let messages = [
        {
          role: "user",
          content: [
            { type: "text", text: argPrompt },
            {
              type: "image_url",
              image_url: { url: imageData.src, detail: "high" },
            },
          ],
        },
      ];

      let resultText = "";
      await requestEngine.chatRequest(messages, false, async function (data) {
        if (data) {
          resultText += data;
        }
      });

      await Asc.Editor.callMethod("EndAction", ["GroupActions"]);
      await Asc.Editor.callMethod("EndAction", [
        "Block",
        "AI (" + requestEngine.modelUI.name + ")",
      ]);

      Asc.scope.text = resultText || "";

      if (!Asc.scope.text.trim()) {
        await insertMessage(
          "⚠ AI request failed (maybe the model cannot handle images)."
        );
        return;
      }
      await insertMessage(Asc.scope.text);
    } catch (e) {
      try {
        await Asc.Editor.callMethod("EndAction", ["GroupActions"]);
        await Asc.Editor.callMethod("EndAction", [
          "Block",
          "AI (describeImage)",
        ]);
      } catch (ee) {
        /* ignore */
      }
      console.error("describeImage error:", e);
      await insertMessage(
        "An unexpected error occurred while describing the image."
      );
    }
  };

  return func;
})();

Step 4: Compile your changes

This is the most important step in the new workflow. The plugin cannot read your new file directly yet. You must run the helper script to merge your new file into the main plugin logic.

1. Open your terminal.
2. Navigate to the helpers directory:

cd .dev/helpers

3. Run the Python build script:

python3 helpers.py

The helpers.py script will scan the word/, cell/, and slide/ folders, find your new describe-image.js, and merged it into the main helpers.js file.

 Step 5: Test your feature

1. Reload the plugin in ONLYOFFICE.
2. Select an image in your document.
3. Ask the AI: “Describe this image” or “Write a caption for this.”

The AI will now recognize your new custom function and execute the logic you just wrote.