On the log details page, I can see the full log content (full prompt and response from the model). But when I curate a dataset and export it (as CSV/JSONL/Drive – the issue is consistent across all options), logs are truncated, and prompts/responses are unusable for further use.
This has been a problem for a couple of months and still isn’t fixed.
Hello @urukalo,
Could you please let us know at what length your prompts or responses are being truncated?
@Rebbanpalli_Naveen Both requests and responses are being trimmed to roughly 205–230 characters
Here is one example of an exported item:
{“request”:{“model”:“models/gemini-3.1-flash-lite-preview”,“contents”:[{“parts”:[{“text”:“Parse Serbian restaurant order into catalog items.\n\nCATALOG (productName [unit], ★\u003dfrequent):\nŠpargla zelena [kg]\nSuncokret Oljušteni Sirovi [kg]\nVlašac [kom,kg]\nLuk Beli [kg,kom] ★\nSusam [kg]\n”}],“role”:“user”}]},“response”:[{“candidates”:[{“content”:{“parts”:[{“text”:“{\n "items": [\n {\n "productName": "Mladi Kupus",\n "quantity": 3,\n "unit": "kom",\n "confidence": 0.9\n },\n {\n "productName": "Paradajz",\n "quantity": 2,\n "uni”}],“role”:“model”}}]}],“turnId”:“v1_*****QtQTQ”,“datasetIds”:[“S_rfaaaRE******PvsudoQ4”],“createTime”:“2026-04-**T17:50:51.158418Z”,“responseStatus”:{},“apiSource”:“API_SOURCE_GENERATE_CONTENT”}
any idea/news about? @Rebbanpalli_Naveen