Softpiper Solutions

Education · Knowledge and data

AI tutor with text and voice

A Socratic tutor grounded in your course material, with a voice mode in many languages, audio lessons and rubric-based grading that teachers review.

  • ClaudeAnthropic
  • GeminiGoogle
  • Typical timeline: 6–12 weeks

The problem

Learners need personalised, multilingual practice and feedback at scale, and teachers don't have the hours to give it.

  • Practice and feedback are limited to class time.
  • Learners in regional languages get less support.
  • Grading short answers takes teachers' evenings.

How we solve it

A tutor that guides learners through problems rather than handing out answers, grounded in your syllabus, with a voice mode, audio lessons and first-pass grading that teachers review.

It plugs into LMS (Moodle, Canvas), Google Classroom, Supabase / Postgres and WhatsApp, so nothing you rely on today is replaced. The architecture below shows each step and which model does the work.

Architecture

From your stack, through the AI layer, to numbers you can track

Each step names the model or system doing the work. Writes to your systems wait for a person to approve them.
AI tutor with text and voice. Your systems (LMS (Moodle, Canvas), Google Classroom, Supabase / Postgres and WhatsApp) connect to Softpiper's AI layer, which runs these steps: Ground in the course (Prompt caching); Text tutor (Claude Sonnet); Voice tutor (Gemini Live API); Audio lessons (Gemini text-to-speech); Rubric grading (Claude Haiku); Nudges and reporting (LMS · WhatsApp). Outcomes are measured as practice sessions per learner, completion rate, assessment score change (before and after) and teacher grading time.

Your stack

  • LMS (Moodle, Canvas)
  • Google Classroom
  • Supabase / Postgres
  • WhatsApp

AI layer

Softpiper AI layer

  • Claude
  • Gemini
  1. Ground in the course, using Prompt caching

    Syllabus, notes and question banks are loaded into a cached prompt and a retrieval index.

  2. Text tutor, using Claude Sonnet

    Learners work through problems with hints and questions rather than answers.

  3. Voice tutor, using Gemini Live API

    Learners talk to the tutor and can interrupt naturally.

  4. Audio lessons, using Gemini text-to-speech

    Lessons and summaries are voiced in the learner's language.

  5. Rubric grading, using Claude Haiku

    Short answers are scored against your rubric; teachers review and adjust.

  6. Nudges and reporting, using LMS · WhatsApp

    Progress syncs to your LMS, and study reminders go out on WhatsApp.

Outcomes you measure

  • Practice sessions per learner
  • Completion rate
  • Assessment score change (before and after)
  • Teacher grading time

Why this model

Claude and Gemini, each where it leads

We choose per task on capability, cost and where your data must stay. Final versions are picked at kickoff.
ClaudeAnthropic

Claude is a patient Socratic tutor that follows your pedagogy and stays within your syllabus, held in a cached prompt so sessions stay fast and inexpensive. A small Claude model grades short answers against a rubric with structured output.

GeminiGoogle

The Gemini Live API runs a natural voice tutor that learners can interrupt, and Gemini's text-to-speech covers 130 languages for audio lessons.

Delivery

What you get, how long it takes, and how we keep it safe

Guardrails

  • Age-appropriate content policies, tested in the evaluation set.
  • Parental consent and children's-data rules, including India's DPDP provisions for minors.
  • The tutor guides; it doesn't hand over answers to graded work.
  • Teachers review grades before they count.

Plus the guardrails on every build: facts from your systems, approvals for writes and money, a human one step away, audit logs and evals. See all

Deliverables

  • Tutor grounded in your syllabus
  • Voice mode and audio lessons
  • Rubric grading with teacher review
  • LMS integration
  • Learning analytics dashboard

Typical timeline

  1. Text tutor MVP6–8 weeks
  2. Voice tutor+3–4 weeks

A planning range for a first production release. We confirm scope, price and dates in a tailored proposal.

KPIs to track

  • Practice sessions per learner
  • Completion rate
  • Assessment score change (before and after)
  • Teacher grading time
  • Languages covered

Goals we agree with you and measure from a baseline. They are what we track, not results we promise.

Related services

The services behind this playbook

More playbooks

Related playbooks

All playbooks

Pip · Softpiper's AI

Ask Pip about this playbook

Ask Pip how this would work for you

Pip is Softpiper's AI. Tell it what you run today and it will walk through how this playbook fits, what we would build first and a tailored price range. Or book a call with the team.

Book a call