gwendoline 0.1.0 → 0.1.2

package/README.md

@@ -32,6 +32,24 @@ echo "Why is the sky blue?" | gwen
 
 cat prompt.md | gwen
 cat prompt.md | gwen --cloud
-cat prompt.md | gwen --model "gpt-oss:120b-cloud"
-cat prompt.md | gwen --model "gpt-oss:120b-cloud" > output.md
+cat prompt.md | gwen --model gpt-oss:120b-cloud
+cat prompt.md | "gwen --model gpt-oss:120b-cloud" > output.md
+cat input.json | gwen --chat > output.json
+```
+
+## Chat Mode Usage
+
+Create a file with input message first or pipe it. Then run with parameter `--chat`.
+In chat mode, Gwendoline is expecting the input to be already a list of chat messages. This must already include at least the message, you want to ask now. The output will be a list of chat messages as well, including the response from LLM.
+
+For example, create file `chat.json` with the content:
+
+```json
+[{ "role": "user", "content": "Why is the sky blue?" }]
+```
+
+Run command:
+
+```sh
+cat chat-input.json | gwendoline --chat --model gpt-oss:120b-cloud > chat-output.json
 ```

package/build/index.js

@@ -6,6 +6,7 @@ const LLM_MODEL_LOCAL = "qwen3:4b";
 const LLM_MODEL_CLOUD = "gpt-oss:120b-cloud";
 const isCloudLLM = argv.includes("--cloud");
 const hasLLMSpecified = argv.includes("--model");
+const isChatMode = argv.includes("--chat");
 let customModelName = "";
 if (hasLLMSpecified) {
     argv.forEach((val, index) => {
@@ -22,8 +23,20 @@ async function main() {
         input += chunk;
     });
     process.stdin.on("end", async () => {
-        const content = await runLLMRequest(input.trim());
-        process.stdout.write(content);
+        if (isChatMode) {
+            try {
+                const inputMessages = JSON.parse(input.trim() || "[]");
+                const content = await runLLMRequest(inputMessages, isChatMode);
+                process.stdout.write(content);
+            }
+            catch (error) {
+                throw Error("Could not parse input of chat messages", error || "");
+            }
+        }
+        else {
+            const content = await runLLMRequest([{ role: "user", content: input.trim() }], isChatMode);
+            process.stdout.write(content);
+        }
     });
     if (process.stdin.isTTY) {
         const rl = readline.createInterface({
@@ -36,7 +49,7 @@ async function main() {
                 process.stdout.write("Bye!");
                 process.exit(1);
             }
-            const content = await runLLMRequest(prompt);
+            const content = await runLLMRequest([{ role: "user", content: prompt }], false);
             process.stdout.write(content);
             process.exit(1);
         });
@@ -46,7 +59,7 @@ main().catch((error) => {
     console.error("Fatal error in main():", error);
     process.exit(1);
 });
-async function runLLMRequest(prompt = "") {
+async function runLLMRequest(messages, returnChat = false) {
     const LLM_MODEL = isCloudLLM ? LLM_MODEL_CLOUD : LLM_MODEL_LOCAL;
     try {
         const ollama = new Ollama({
@@ -59,8 +72,16 @@ async function runLLMRequest(prompt = "") {
         });
         const response = await ollama.chat({
             model: customModelName || LLM_MODEL,
-            messages: [{ role: "user", content: prompt }],
+            messages,
         });
+        if (returnChat) {
+            messages.push({
+                role: "assistant",
+                content: response.message.content,
+            });
+            const messagesStr = JSON.stringify(messages);
+            return messagesStr.trim();
+        }
         return response.message.content;
     }
     catch (e) {

package/package.json

@@ -1,7 +1,13 @@
 {
   "name": "gwendoline",
-  "version": "0.1.0",
-  "description": "",
+  "version": "0.1.2",
+  "description": "CLI based tool for interacting with language models directly from terminal",
+  "keywords": [
+    "cli",
+    "llm",
+    "ollama",
+    "chat"
+  ],
   "type": "module",
   "bin": {
     "gwen": "./build/index.js",