1. Chatting with Sasini AI
The AI is built for deep, meaningful conversations. It understands context, remembers your current session, and provides highly accurate, Sasini-approved advice.Ask Anything
Simply type your question into the chat box. You have a massive 1,000-character limit per message, allowing you to explain your farm’s exact situation in deep detail.
Get Instant Answers
The AI processes complex algorithms in milliseconds. While it is trained to answer general world knowledge, its primary directive is to prioritize Sasini-related protocols, crop management, and logistical help.

Backend Security: Gemini API & Vector Embeddings
Backend Security: Gemini API & Vector Embeddings
Under the Hood: We don’t just use a generic AI. User queries are routed through a secure API tunnel to the Google Gemini Pro engine which powers the Sasini AI. Before answering, the system queries the Database containing massive databets of Sasini PDF manuals and datasets, injecting that specific knowledge into the AI’s context window.
Backend Security: Hard Delete Protocol
Backend Security: Hard Delete Protocol
Under the Hood: When a user taps ‘Delete Chat’, the frontend sends a
DELETE HTTP request to the server. We enforce a “Hard Delete” at the database level (e.g., MongoDB or PostgreSQL), completely dropping the user’s session ID thread. No ghost data is kept.2. Real-Time Image Diagnosis
The crown jewel of Sasini AI is its multi-modal vision capability. It can “see” what you see.Upload a Photo
Tap the Attachment/Camera icon next to the chat bar to take a live photo of a pest, a diseased leaf, or even a broken farm tool.
Add Context
Type a brief message along with the image, such as “What bug is this eating my macadamia?” and hit send.
Lighting Matters: For the most accurate AI diagnosis, ensure your photo is taken in bright daylight and the camera is fully focused on the affected area.

Backend Security: Base64 Encoding & Vision Processing
Backend Security: Base64 Encoding & Vision Processing
Under the Hood: Images uploaded to the AI are converted into a secure Base64 payload or temporarily hosted on a bucket before being passed to the Sasini AI. The image is purged from the temporary processing cache immediately after the inference is returned to save server costs.