Argus- It never forgets

Data is gold, but stays passive in chats. Argus is a proactive Gemini 3 agent that connects past WhatsApp chats to present browsing, injecting context when you need it. Argus: It never forgets

Comment

Inspiration

We were inspired by a simple irony: we live in the age of 'Big Data,' yet our most valuable personal data our conversations is passive. It sits in a 'chat grave' while we make decisions in a browser. We wanted to turn that 'Passive Gold' into 'Active Intelligence PS : I forgot to cancel my Netflix Subscription and I almost every time miss quiz

What it does

Argus is a proactive intelligence layer that sits between your private communications and your active web browsing. It eliminates the "Context Gap."

  1. Passive Data Ingestion The Input: Argus uses the Evolution API to securely listen to your WhatsApp messages. The Storage: Instead of letting messages disappear in a scroll, it stores them in a Postgres database, turning "chat noise" into a structured personal knowledge base.

  2. Intent-Based Sensing (The Event-Trigger) The Sensor: As you browse the web (Amazon, Nykaa, College Portals, etc.), the Argus Chrome Extension monitors your current URL and page content. The Trigger: It doesn't run 24/7; it triggers only when your browsing suggests an action that might have a corresponding "memory" in your chats.

  3. Proactive Recall (Powered by Gemini 3) The Brain: Argus sends your current browsing context to Gemini 3. The Match: Gemini queries your past WhatsApp history to find relevant links, dates, or recommendations. The Result: If it finds a match, it doesn't wait for you to search it injects a smart notification directly into your current webpage.

How we built it

We built a multi-layered system that bridges real-time messaging with live browser state using an event-driven architecture.

  1. Data Ingestion & Pipeline (The Ears) Evolution API: We used the Evolution API to create a bridge to WhatsApp. This allows us to receive webhooks for incoming messages in real-time. Vector Ready Storage: Messages are processed and stored in a Postgres database. We structured the data to ensure that metadata (sender, timestamp, intent) is preserved for high-accuracy retrieval.

  2. Intent Sensing (The Eyes) Chrome Extension (Manifest V3): The frontend is a lightweight extension. We avoided the "passive icon" trap by using Background Service Workers to monitor navigation events. Constrained Triggers: To ensure privacy and performance, the extension uses a URL Listener. When a user visits a high-intent site (e.g., shopping, university portals, subscriptions), it triggers a background handshake with our backend.

  3. The Intelligence Layer (The Brain) Gemini 3 Integration: This is the core of Argus. We leverage Gemini 3 for its Long-Context Window and Reasoning Capabilities. Contextual Fetching: Instead of a basic keyword search, we send the current browsing URL/Context to Gemini. Gemini then "reasons" through the stored WhatsApp history to find semantically relevant info (e.g., matching a product recommendation to a specific Nykaa page).

  4. Zero-Click Injection (The Voice) Content Script Injection: Once a match is found, the backend sends a payload back to the extension. DOM Manipulation: We use content scripts to dynamically inject a custom Shadow DOM element into the current webpage. This ensures our "Smart Recall" UI doesn't break the original website's styling and feels like a native part of the experience.

Key Technical Challenges Solved: Contextual Relevance: Ensuring the AI doesn't "spam" the user by only showing notifications when the intent score is high. Real-time Latency: Optimizing the bridge between the Evolution API and the Browser Extension to ensure the "Recall" happens as the page loads.

Accomplishments that we're proud of

Beyond the code, we are proud of building a tool that solves the "Forgetfulness Tax." In our internal testing, Argus successfully caught three real-world missed deadlines (a quiz, a bill payment, and a subscription trial) that would have otherwise cost us time and money.

What we learned

Gemini 3 isn't just a faster chatbot; it’s a reasoning engine. It moved Argus from being a "search tool" to a proactive agent that understands the relationship between a user's past life and their current digital environment.

Gemini 3 features used-

  1. GEMINI FLASH (Fast, Low-Cost)

    • Model: gemini-3.0-flash-preview
    • Use: Event extraction from WhatsApp
    • Why: Low latency, good accuracy

  2. JSON MODE (Structured Output)

    • Feature: response_mime_type = "application/json"
    • Use: Generate popup blueprints
    • Why: Consistent UI rendering

  3. LARGE CONTEXT (128K tokens)

    • Feature: Standard Gemini Flash context
    • Use: Process multiple messages
    • Why: Understand conversation history

  4. SEMANTIC UNDERSTANDING

    • Feature: Built-in reasoning
    • Use: Intent detection, context matching
    • Why: Distinguish actions from mentions

  5. MULTI-TURN CONVERSATIONS

    • Feature: Conversation history
    • Use: AI chat sidebar
    • Why: Contextual Q&A

What's next for Argus- It never forgets

Share this project:

Updates