Verwendung
VerwendungDebugging von API-Anfragen

Debugging von API-Anfragen

Um die an Übersetzungsanbieter (wie ChatGPT, Claude oder Google Translate) gesendeten Anfragen und deren Antworten zu debuggen, kannst du den Schweregrad 🔵 Info in den Einstellungen von Logs & Notifications aktivieren.

Info-Schweregrad in den Einstellungen von Logs & Notifications aktivieren
Info-Schweregrad in den Einstellungen von Logs & Notifications aktivieren

Zugriff auf die API-Anfragen-Logs

Die Logs werden unter den Einträgen api-requests gespeichert.

KI-Anfragen in den Logs
KI-Anfragen in den Logs

Was protokolliert wird

Für KI-Anbieter enthält der Eintrag api-requests detaillierte Informationen über:

  • Den an den KI-Anbieter gesendeten Prompt
  • Die vollständige empfangene Antwort
  • Etwaige Fehler oder Probleme während der Kommunikation
  • Das verwendete Modell
  • Die Anzahl der verwendeten Tokens
Detail einer KI-Anfrage in den Logs
Detail einer KI-Anfrage in den Logs

Zum Beispiel zeigt der folgende JSON „Additional context" die Details einer an Claude gesendeten Anfrage und deren Antwort:

{
  "request": {
    "model": "claude-3-7-sonnet-latest",
    "max_tokens": 128000,
    "messages": [
      {
        "role": "user",
        "content": "You are a language translator.\r\n\r\nI'm working on internationalizing my application.\r\n\r\nI've created a JSON with sentences in English. Please translate the sentences to Spanish from Argentina.\r\n\r\nIf a sentence contains HTML, do not translate inside the HTML tags.\n\nReturn ONLY a JSON object with a single key \"translations\" containing an array of translated strings.\nDo not include any explanations, markdown formatting, or code blocks.\nThe response must be a valid JSON object starting with { and ending with }.\n\nThis is the JSON:\n\n[\"HTTP caching improves performance\",\"Categories Block\",\"Latest Posts Block\",\"Democratizing publishing\"]"
      }
    ]
  },
  "response": {
    "id": "msg_01EC2Qp2fMw6VQY8tGESFrh2",
    "type": "message",
    "role": "assistant",
    "model": "claude-3-7-sonnet-20250219",
    "content": [
      {
        "type": "text",
        "text": "{\n  \"translations\": [\n  \"El caché HTTP mejora el rendimiento\",\n  \"Bloque de categorías\",\n  \"Bloque de entradas recientes\",\n  \"Democratizando la publicación\"\n  ]\n}"
      }
    ],
    "stop_reason": "end_turn",
    "stop_sequence": null,
    "usage": {
      "input_tokens": 138,
      "cache_creation_input_tokens": 0,
      "cache_read_input_tokens": 0,
      "output_tokens": 61,
      "service_tier": "standard"
    }
  }
}