@ducci/jarvis 1.0.40 → 1.0.41

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ducci/jarvis",
-  "version": "1.0.40",
+  "version": "1.0.41",
   "description": "A fully automated agent system that lives on a server.",
   "main": "./src/index.js",
   "type": "module",

package/src/server/agent.js

@@ -50,6 +50,12 @@ async function callModel(client, model, messages, tools) {
   return await client.chat.completions.create(params);
 }
 
+function isImageUnsupportedError(apiErrors) {
+  if (!apiErrors) return false;
+  return [apiErrors.primary?.message, apiErrors.fallback?.message]
+    .some(m => m?.toLowerCase().includes('image input'));
+}
+
 function extractApiError(err, model) {
   return {
     model,
@@ -640,6 +646,9 @@ async function _runHandleChat(config, sessionId, userMessage, attachments = [])
         // windows). The synthetic note is sufficient context; tool results are preserved
         // in the JSONL log and accessible via read_session_log.
         if (finalStatus === 'model_error' || finalStatus === 'format_error') {
+          if (finalStatus === 'model_error' && isImageUnsupportedError(run.errorDetail)) {
+            finalResponse = 'This model does not support image input. Please switch to a multimodal model (e.g. claude-3.5-sonnet, gpt-4o) in settings.';
+          }
           session.messages.splice(runStartIndex, session.messages.length - runStartIndex);
           const errorDetail = run.errorDetail ? ` Error detail: ${JSON.stringify(run.errorDetail)}` : '';
           session.messages.push({