Voice Mentor Chatbot for Students
Budget: ₹400 – ₹750 INR
I need a generative-AI developer to build a small-scale MVP of a multimodal chatbot that can act as a daily mentor for students preparing for Chartered Accountancy. The first release must live on a website and support natural, two-way voice conversations—speech-to-text on the way in, text-to-speech on the way out—so learners can talk to it as if they were speaking to a tutor.
Core goals
• Accurate CA guidance: the bot should answer syllabus-level questions, explain tricky concepts, and suggest study plans.
• Fluid voice exchange: latency below two seconds using a stack such as Whisper / Web Speech API for recognition and a neural TTS engine for replies.
• Continual engagement: greet students each day, track brief study logs, and offer tailored prompts to keep them on pace.
What I already have
• A basic list of common CA queries and official reference links.
• Wireframes for a minimal web UI (chat window + mic button).
• A modest hosting plan on Vercel.
Where I need your expertise
• Selecting and wiring up an LLM (GPT-4, Claude, etc.) with proper prompt engineering for CA content.
• Integrating voice I/O and streaming the responses.
• Building the lightweight frontend (React preferred, but plain JS is fine) and a simple Node or Python backend to handle auth, session memory, and API calls.
• Optimising costs so the solution stays affordable until I can scale.
Deliverables
1. Working web prototype deployed to my domain.
2. Clean, documented code in GitHub.
3. Quick-start guide for future feature additions.
4. One short hand-off call to walk me through deployment and cost controls.
Acceptance criteria
– Users can press a button, speak a CA question, and hear a voiced response within two seconds.
– Answers reflect the latest ICAI material and include source citations.
– Daily check-in prompt fires automatically after 24 hours of inactivity.
– All third-party services remain within the free or lowest-tier plans.
If you have prior experience with conversational AI, voice pipelines, or education tech, let’s talk—I’m aiming to start immediately and keep the first milestone tight so we stay within budget.
Core goals
• Accurate CA guidance: the bot should answer syllabus-level questions, explain tricky concepts, and suggest study plans.
• Fluid voice exchange: latency below two seconds using a stack such as Whisper / Web Speech API for recognition and a neural TTS engine for replies.
• Continual engagement: greet students each day, track brief study logs, and offer tailored prompts to keep them on pace.
What I already have
• A basic list of common CA queries and official reference links.
• Wireframes for a minimal web UI (chat window + mic button).
• A modest hosting plan on Vercel.
Where I need your expertise
• Selecting and wiring up an LLM (GPT-4, Claude, etc.) with proper prompt engineering for CA content.
• Integrating voice I/O and streaming the responses.
• Building the lightweight frontend (React preferred, but plain JS is fine) and a simple Node or Python backend to handle auth, session memory, and API calls.
• Optimising costs so the solution stays affordable until I can scale.
Deliverables
1. Working web prototype deployed to my domain.
2. Clean, documented code in GitHub.
3. Quick-start guide for future feature additions.
4. One short hand-off call to walk me through deployment and cost controls.
Acceptance criteria
– Users can press a button, speak a CA question, and hear a voiced response within two seconds.
– Answers reflect the latest ICAI material and include source citations.
– Daily check-in prompt fires automatically after 24 hours of inactivity.
– All third-party services remain within the free or lowest-tier plans.
If you have prior experience with conversational AI, voice pipelines, or education tech, let’s talk—I’m aiming to start immediately and keep the first milestone tight so we stay within budget.