Personal-AI-Clone

Gemini 3.1 Pro Preview Ran for 151s

summarize_auto Action history chevron_right Here are key actions taken for the app:

edit Edited 4 files metadata.json check_circle src/lib/gemini.ts check_circle src/lib/utils.ts check_circle src/App.tsx check_circle build Built I have finished building out your AI Personal Clone Assistant based on your PRO PROMPT. Let me walk you through the architecture and features: Features and Architecture Clone Persona Engine: You are given a prompt setup UI right alongside your workspace. You can edit the user personality (traits), long-term goals/motivations, and past memories/decisions, replicating the context a standard Vector DB implementation would supply dynamically over time. Decision Engine in Chat: On the right, when you ask your clone a question, it forwards the injected persona along with your prompt directly effectively simulating the prompt chain logic you described. Model Output Configuration: The app connects to the gemini-2.5-flash model and is wired up to strictly output your specific JSON specification { "response_as_user", "reasoning", "confidence" } using the responseSchema property provided by the native @google/genai TypeScript API. Test it out You can immediately try providing your simulated profile details on the left, then ask the AI decisions from the bottom right like: "Should I invest in a risky startup or save money?" The Clone will reply contextually, simulating your behavior, and attach the reasoning block with evaluating the confidence of its decision! flag Checkpoint

Built With

• 1.9%
• 98.0%
• css
• html
• typescript