An assistant that speaks Thai and remembers your farm
NaLog Agent turns sensor data and farm history into short, plain-language advice, and prepares pump actions that the farmer approves or declines.
NaLog Agent is an AI assistant for rice farmers, deployed on Alibaba Cloud. It answers in Thai or English, remembers each farmer’s paddies and conditions across sessions, and proposes irrigation actions that require the farmer’s explicit approval before execution.
Thai and English
Answers in the farmer’s language, using live data from their own paddies in NaLog.
Long-term memory
Remembers facts about each farmer and paddy across sessions and seasons. When new information contradicts an old fact, the old one is superseded rather than kept alongside it.
Farmer approval, always
The agent cannot switch a pump. It creates a proposal; only the farmer’s approval sends the command.
How it works
- Reads live sensor data from your paddies and knows your farm's history.
- If something new contradicts older information, it updates automatically — no stale advice.
- Can read photos of your field to help assess conditions.
- Reacts to sensor alerts without waiting for you to ask.
NaLog Agent received an honourable mention with a monetary award at an Alibaba Cloud / Qwen hackathon (Track 1: MemoryAgent) in 2026, recognised for technical depth and blog posting.
See it working
A demo of NaLog Agent in action. The founder’s write-up explains the design:NaLog Agent: a Qwen memory agent demo.