Inspiration Hamari inspiration school ke un dino se aati hai jab itihaas (History) padhna sirf dates aur naam yaad karne tak seemit tha. Textbooks me likhi kahaniyan bejaan lagti thi. Humne socha, "Kya ho agar hum in itihaas ke panno se characters ko bahar nikal kar unse seedhe baat kar paayein?" Isi soch ne Itihaas Sakshi ko janm diya—ek aisa platform jahan itihaas ek subject nahi, balki ek anubhav hai.

What it does Itihaas Sakshi ek AI-powered web application hai jo users ko Mahatma Gandhi, Rani Lakshmibai, aur Leonardo da Vinci jaise mahaan aitihasik patron se real-time mein baat karne ka mauka deta hai.

Select & Chat: User apni pasand ka historical figure chunta hai.

Converse in Real-Time: Woh unse kuch bhi pooch sakte hain—unke jeevan, unke samay, ya unki uplabdhiyon ke baare mein.

See History: Baat-cheet ke dauran, AI unke jawab ko behtar samjhane ke liye relevant images generate karta hai. Jaise, Einstein se relativity par baat karte hue ek diagram ya Rani Lakshmibai se unke kile ki tasveer.

Hear History: Har jawab ko user uss character ki awaaz mein sun sakta hai, jisse experience aur bhi immersive ho jaata hai.

Never Get Stuck: Agar user ko samajh na aaye ki kya poochna hai, toh ek "Suggest a Question" feature unhein creative sawal sujhao deta hai.

How we built it Yeh poora project Google Gemini API ki shakti par banaya gaya hai, jise humne simple frontend technologies ke saath integrate kiya hai.

Frontend: HTML, Tailwind CSS, Vanilla JavaScript.

Core AI Logic (The Gemini Trinity):

Conversational AI: Humne gemini-2.5-flash-preview-05-20 ka istemal kiya. Advanced system prompts ke zariye humne AI ko har character ki personality, knowledge, aur baat karne ka tarika sikhaya.

Image Generation: gemini-2.5-flash-image-preview model on-the-fly contextual images generate karta hai, jo chat mein dikhayi jaati hain.

Text-to-Speech (TTS): gemini-2.5-flash-preview-tts model AI ke text responses ko ek realistic awaaz deta hai, jiske liye humne har character ke liye ek unique voice chuni hai.

Challenges we ran into Effective Prompt Engineering: AI ko hamesha character mein rakhna aur sahi tone maintain karwana shuruaat mein mushkil tha. Iske liye humne system prompts ko kai baar refine kiya.

Asynchronous API Handling: Ek hi samay par text, image, aur audio ke liye API calls ko manage karna aur UI ko smooth rakhna ek challenge tha.

Client-Side Audio Conversion: Gemini TTS API raw PCM audio data return karti hai. Is base64 encoded data ko browser mein chalane layak .wav file mein convert karne ke liye humein custom JavaScript functions likhne pade.

Accomplishments that we're proud of Humne ek multi-modal learning experience (text, image, audio) banaya hai, aur yeh sab sirf ek API suite (Gemini) ka istemal karke kiya hai.

Application ka user interface behadd intuitive aur engaging hai, jo learning ko ek boring kaam se ek mazedaar activity bana deta hai.

Bina kisi traditional backend ke, humne ek fully functional, dynamic, aur powerful web app banayi hai.

What we learned Is project ke dauran humne Gemini API ki versatility ko gehrai se jaana. Humne seekha ki kaise effective prompt engineering se AI ke output ko control kiya ja sakta hai. Saath hi, client-side par audio data manipulation aur multiple asynchronous operations ko gracefully handle karna hamare liye ek badi learning thi.

What's next for Itihaas Sakshi Hamara vision Itihaas Sakshi ko ek global educational tool banana hai.

More Characters: Alag-alag deshon aur samay ke historical figures ko add karna.

Multi-Language Support: Users ko apni native language mein baat karne ki suvidha dena.

Curriculum Integration: Schools ke liye specific learning modules banana (e.g., "Mughal Empire," "World War II").

Group Conversations: Ek aisi feature jahan user do ya do se zyada historical figures ko ek saath ek "virtual room" mein laakar unse kisi topic par debate karwa sake.

Share this project:

Updates