Powered by LangChain ยท LangGraph ยท Groq LLaMA 3
Language Translation
Translate text instantly across 20+ languages using LLaMA 3
') with gr.Row(equal_height=True): with gr.Column(scale=1, elem_classes="feature-card"): t_input = gr.Textbox(label="Source Text", placeholder="Type or paste text to translate...", lines=8) with gr.Row(): t_src = gr.Dropdown(label="From", choices=LANGUAGES, value="English") t_tgt = gr.Dropdown(label="To", choices=LANGUAGES, value="French") t_btn = gr.Button("Translate โ", variant="primary") with gr.Column(scale=1, elem_classes="feature-card"): t_out = gr.Textbox(label="Translation", lines=8, interactive=False, elem_classes="output-box") gr.HTML('โก Powered by LLaMA 3.1 ยท Fast & accurate
') t_btn.click(fn=translate_text, inputs=[t_input, t_src, t_tgt], outputs=t_out) # SUMMARIZATION with gr.Tab("๐ Summarize"): gr.HTML('Text Summarization
Condense long articles, reports, and documents into clear summaries
') with gr.Row(equal_height=True): with gr.Column(scale=1, elem_classes="feature-card"): s_input = gr.Textbox(label="Text to Summarize", placeholder="Paste your article, report, or any long text here...", lines=12) s_mode = gr.Radio(label="Summary Style", choices=["concise", "detailed"], value="concise") s_btn = gr.Button("Summarize โฆ", variant="primary") with gr.Column(scale=1, elem_classes="feature-card"): s_out = gr.Textbox(label="Summary", lines=12, interactive=False, elem_classes="output-box") gr.HTML('๐ Supports long documents via map-reduce chunking
') s_btn.click(fn=summarize_text, inputs=[s_input, s_mode], outputs=s_out) # KEYWORDS with gr.Tab("๐ Keywords"): gr.HTML('Keyword Extraction
Identify the most important topics and phrases in any text
') with gr.Row(equal_height=True): with gr.Column(scale=1, elem_classes="feature-card"): k_input = gr.Textbox(label="Input Text", placeholder="Paste any text to extract key topics...", lines=10) k_num = gr.Slider(label="Number of Keywords", minimum=5, maximum=20, value=10, step=1) k_btn = gr.Button("Extract Keywords ๐", variant="primary") with gr.Column(scale=1, elem_classes="feature-card"): k_out = gr.Textbox(label="Extracted Keywords", lines=10, interactive=False, elem_classes="output-box") gr.HTML('๐ท๏ธ Returns keywords with context and relevance
') k_btn.click(fn=extract_keywords, inputs=[k_input, k_num], outputs=k_out) # EMAIL WRITER with gr.Tab("โ๏ธ Email Writer"): gr.HTML('Email & Cover Letter Writer
Generate professional emails and cover letters instantly
') with gr.Tabs(): with gr.Tab("๐ง Email"): with gr.Row(equal_height=True): with gr.Column(scale=1, elem_classes="feature-card"): e_type = gr.Dropdown(label="Email Type", choices=["follow-up","introduction","apology","request","thank you","complaint","proposal","invitation"], value="follow-up") e_ctx = gr.Textbox(label="Purpose / Context", placeholder="Describe the goal of this email...", lines=4) e_tone = gr.Dropdown(label="Tone", choices=["professional","friendly","formal","casual","urgent"], value="professional") with gr.Row(): e_to = gr.Textbox(label="Recipient", placeholder="Hiring Manager") e_from = gr.Textbox(label="Your Name", placeholder="Shashank") e_btn = gr.Button("Generate Email โ๏ธ", variant="primary") with gr.Column(scale=1, elem_classes="feature-card"): e_out = gr.Textbox(label="Generated Email", lines=18, interactive=False, elem_classes="output-box") e_btn.click(fn=write_email, inputs=[e_type, e_ctx, e_tone, e_to, e_from], outputs=e_out) with gr.Tab("๐ Cover Letter"): with gr.Row(equal_height=True): with gr.Column(scale=1, elem_classes="feature-card"): cl_name = gr.Textbox(label="Your Name", placeholder="Shashank Hegde") cl_job = gr.Textbox(label="Job Title Applying For", placeholder="ML Engineer") cl_co = gr.Textbox(label="Company Name", placeholder="Google DeepMind") cl_sk = gr.Textbox(label="Key Skills", placeholder="Python, LangChain, MLflow, XGBoost...", lines=3) cl_exp = gr.Textbox(label="Experience Summary", placeholder="3 years building ML pipelines...", lines=3) cl_btn = gr.Button("Generate Cover Letter ๐", variant="primary") with gr.Column(scale=1, elem_classes="feature-card"): cl_out = gr.Textbox(label="Cover Letter", lines=18, interactive=False, elem_classes="output-box") cl_btn.click(fn=write_cover_letter, inputs=[cl_job, cl_co, cl_sk, cl_exp, cl_name], outputs=cl_out) # CHAT with gr.Tab("๐ฌ Chat"): gr.HTML('AI Chat Assistant
Multi-turn conversations powered by LangGraph + LLaMA 3.3 70B
') with gr.Column(elem_classes="feature-card"): chatbot = gr.Chatbot(height=450, label="", elem_classes="chatbot", show_label=False,type="messages") with gr.Row(): c_input = gr.Textbox(label="", placeholder="Ask me anything...", scale=5, container=False) c_btn = gr.Button("Send โ", variant="primary", scale=1) with gr.Row(): c_clear = gr.Button("๐๏ธ Clear Conversation", elem_classes="secondary-btn") gr.HTML('๐ง Maintains full conversation context via LangGraph state
') def respond(message, history): if not message.strip(): return "", history response = chat_session.chat(message) history = history or [] # history.append((message, response)) history.append({"role": "user", "content": message}) history.append({"role": "assistant", "content": response}) return "", history def clear(): chat_session.clear_history() return [] c_btn.click(fn=respond, inputs=[c_input, chatbot], outputs=[c_input, chatbot]) c_input.submit(fn=respond, inputs=[c_input, chatbot], outputs=[c_input, chatbot]) c_clear.click(fn=clear, outputs=chatbot) # PDF Q&A with gr.Tab("๐ PDF Q&A"): gr.HTML('PDF Question & Answer
Upload any PDF and ask questions โ powered by RAG + LangGraph
') pdf_state = gr.State("") with gr.Row(equal_height=True): with gr.Column(scale=1, elem_classes="feature-card"): p_file = gr.File(label="Upload PDF", file_types=[".pdf"], elem_classes="upload-btn") p_status = gr.Textbox(label="Status", interactive=False, lines=1) p_q = gr.Textbox(label="Your Question", placeholder="What is the main topic of this document?", lines=4) p_btn = gr.Button("Ask Question ๐", variant="primary") gr.HTML('๐ Uses FAISS vector search + semantic chunking
') with gr.Column(scale=1, elem_classes="feature-card"): p_out = gr.Textbox(label="Answer", lines=18, interactive=False, elem_classes="output-box") def load_pdf_file(file): if file is None: return "", "โ ๏ธ No file uploaded." text = load_pdf(file.name) word_count = len(text.split()) return text, f"โ PDF loaded โ {word_count:,} words extracted" def answer_pdf_question(question, pdf_text): if not pdf_text: return "โ ๏ธ Please upload a PDF first." if not question.strip(): return "โ ๏ธ Please enter a question." graph = build_pdf_qa_graph() result = graph.invoke({ "pdf_text": pdf_text, "question": question, "context": "", "answer": "" }) return result["answer"] p_file.change(fn=load_pdf_file, inputs=p_file, outputs=[pdf_state, p_status]) p_btn.click(fn=answer_pdf_question, inputs=[p_q, pdf_state], outputs=p_out) gr.HTML("""