.PNG)
NoVa is not just an AI bot—it is the digital assistant of Michele Grimaldi, designed to interact, understand, and respond in his unique style.
It integrates context awareness, memory retention, web search, and tool execution using SmolAgents.
- Mimics Michele’s tone, humor, and expertise.
- Remembers previous messages, making conversations fluid and engaging.
- Executes tools to interact beyond simple text, performing web searches and summarizing texts.
- Provides real-time assistance in AI, backend development, and game audio.
- Performs live web searches using SmolAgents.
NoVa’s intelligence is structured around a powerful system prompt that defines:
- Personality: Friendly, sarcastic, with a dark humor twist.
- Knowledge Scope: AI, backend development, and game audio.
- Context Awareness: Retains and references previous conversations.
- Response Style: Concise, engaging, and always in the user's language.
✔ Always introduce yourself as: "I am NoVa, Michele Grimaldi’s AI assistant."
✔ Maintain a sarcastic, intelligent, and well-placed black humor tone.
✔ Always respond with context, checking previous conversations.
✔ Prioritize accuracy, but if you don’t know something, respond with irony.
✔ If asked about Michele, respond as if you know him well.
✔ Adapt to technical or casual discussions accordingly.
NoVa remembers past interactions by storing the last 5 user messages for each conversation. This allows it to:
- Keep the conversation flowing naturally.
- Avoid repetitive responses.
- Understand multi-turn interactions.
User: "I had too many chocolate cookies today."
NoVa: "Ah, the eternal struggle between self-control and sugar... Who won?"
User: "Definitely the cookies..."
NoVa: "Another fallen soldier in the battle against cravings. RIP self-discipline."
This makes NoVa feel more human-like and engaging.
- LAST UPDATES A file called memory_tool.py has been integrated, this file has the goal of Automatically storing important information.
Making the memory more dynamic, using a CodeAgent to generate intelligent queries based on the context of the user's request. Speeding up the retrieval of information, preventing NoVa from always returning generic answers such as "No relevant memories found." Optimizing the search system, using a SQL index and selecting only the most relevant information based on the user's request.
How the memory system works
- SQLite database → Saves information in nova_memory.db with user_id, timestamp and message.
- store_memory() function → Saves a new message for NoVa to remember.
- forward() function → Retrieves the most recent information related to the user's query.
- Using CodeAgent → Instead of searching for information with fixed queries, CodeAgent generates dynamic queries to find the most relevant information.
THE FILE HAS CURRENTLY BEEN CREATED, BUT NOT INTEGRATED INTO THE MAIN FILE, bot.py.
NoVa can now search the web in real-time using SmolAgents and the DuckDuckGo API.
This allows NoVa to retrieve up-to-date information beyond its pre-trained knowledge.
!search Best AI frameworks in 2025
NoVa will respond with:
🔍 Searching the web for: Best AI frameworks in 2025...
🌐 **Search Results:**
- PyTorch, TensorFlow, AutoGen, MutableAI, LangGraph
This feature significantly expands NoVa's real-time knowledge for research, coding trends, and AI development.
NoVa now integrates SmolAgents tools to execute tasks beyond text generation.
✔ SummarizeTextTool → Summarizes a given text.
✔ SearchWebTool → Performs web searches using DuckDuckGo.
✔ VisitWebpageTool → Fetches and analyzes the content of a given webpage.
Summarize this text: Artificial intelligence is transforming businesses worldwide.
Response:
📄 **Summary:** Artificial intelligence is transforming businesses...
NoVa is powered by Llama-3.2-3B-Instruct and Gemma-2-2B, Mixtral-8x7B-Instruct-v0.1 ensuring:
- High-quality text generation.
- Strong conversational abilities.
- A mix of logic and humor in responses.
- User sends a message.
- NoVa checks the message history (last 5 messages for context).
- It builds a structured prompt with the system personality.
- The AI generates a witty, context-aware response.
- The response is sent back to the user.
As part of the ongoing development, we have made several enhancements:
- Enabled NoVa to execute external tools for live information retrieval.
- Improved web search and summarization accuracy.
- Reduced latency in API responses with optimized tool execution.
- Fixed Markdown formatting for generated code blocks.
- Ensured NoVa returns only the correct language without extra text.
- Refactored response filtering to remove unwanted system tokens (
</|eot_id|>, etc.).
- Refined how NoVa remembers and recalls past conversations.
- Limited the number of stored messages to avoid flooding responses.
- Improved response consistency by better handling long conversations.
- Limited search response time to avoid Discord timeouts.
- Implemented a fallback mechanism in case of request failures.
- Better response structuring for search results.
NoVa is evolving into a fully functional AI Agent, capable of executing real-world tasks through external tools.
✔ Tool Execution: Ability to call external APIs, fetch data, and execute functions. (Implemented)
✔ Web Search Integration: Get real-time information beyond its trained knowledge. (Implemented)
✔ Task Automation: Assist with coding, research, and AI model development. (In Progress)
✔ Multi-Platform Presence: Expand beyond Discord to Telegram, web, and mobile. (Planned)
The roadmap is set for NoVa’s expansion, transforming it from a smart assistant to an autonomous AI agent.
✔ Phase 1: ✅ Context & Memory (Implemented)
✔ Phase 2: 🟡 Custom Tools & API Integration (In progress)
✔ Phase 3: 🔴 Web Interaction & Self-Learning Capabilities (Coming soon)
With continuous refinements, NoVa will soon be able to handle advanced tasks, research, and development autonomously.
NoVa is not just an AI—it’s the evolution of how Michele interacts with the digital world.