When life feels overwhelming,
find a moment
of clarity.

01 · BUILD

A small, achievable first version

Team size: one. Manya Mahajan leads product direction, user research and campus outreach, using AI coding tools to build the prototype. There are no dedicated frontend, backend or AI/ML engineers on the team. Technical review and implementation support for production integrations remain open dependencies.

Current stack: React, TypeScript and Vinext, with a server-side waitlist endpoint and Cloudflare D1 storage. The current preview uses predefined responses.

Planned personalisation: OpenAI API, called only from the server. Retrieve a relevant approved verse first, then ask the model to explain it and suggest a reflection. Current status: account not created; billing and integration not started. Owner: Manya Mahajan. Target: create the account, activate billing and complete a server-side test by the end of week 1. By the end of week 2, test the complete concern → approved verse → explanation flow. If access is unavailable, retain clearly labelled predefined previews and postpone paid personalisation; a sample response will never be presented as live AI.

Planned accounts: Supabase Auth for email-link sign-in and Razorpay hosted checkout for payments are the proposed managed services; neither will be built from scratch. Current status: neither account has been created; onboarding and integration have not started. Owner: Manya Mahajan. Target: create accounts and submit onboarding requirements in week 1, then complete integration and end-to-end tests by the end of week 2, subject to provider approval. If approval is delayed, keep the free preview available and do not collect payment or promise paid access. Before charging users, test sign-in, payment verification, paid access, failed payments and cancellation. Test-mode checkout will not be described as live payments.

02 · VERSE QUALITY

Begin with 30 carefully checked verses

Source approach: Manually curate an initial collection of 30 verses covering uncertainty, comparison, disappointment and action. Store the chapter, verse number, source link, context, topic tags and an original plain-language paraphrase.

The sample references link to Holy Bhagavad Gita and Vedabase. They are reference sources, not a licensed database or a secured API partnership. Manya reports that a reviewer has agreed to help. The reviewer’s name, qualifications, written agreement and review schedule have not yet been documented here; this is not a claim of completed review or institutional endorsement. Owner: Manya Mahajan. Target: document reviewer details and availability by the end of week 1, then complete the 30-verse review and document any required reuse permissions by the end of week 2. Content permissions remain unconfirmed. A reviewer’s agreement does not grant translation reuse rights. Original paraphrases also require accuracy review.

Release check: For every record, retain its source, permission evidence, reviewer, review date and approval status. Check each reference and paraphrase against its source, obtain a second review, and restrict generation to approved records. If rights or review cannot be secured, exclude the affected material and delay the paid pilot. Label AI interpretations separately. If no suitable verse is available, say so instead of inventing a teaching.

03 · FIRST CUSTOMERS

Start with our campus community

Reachable channel: Founder-led outreach through Manya’s Masters’ Union student network and class WhatsApp groups, with group-admin permission.

Proposed eight-week funnel: Invite 50 students to try the experience, aim for 20 trial users, and seek 5–10 paying users by week eight. These are targets to test, not existing traction.

Offer a seven-day trial and test a founding price of ₹99/month for up to 30 guided exchanges. Show the price before the trial; request explicit payment consent. Measure trial completion, paid conversion and cost per conversation before setting final pricing.

04 · DELIVERY

Clear milestones, visible dependencies

Week 1: Manya owns account creation, API billing activation and testing, provider onboarding, and documenting the reviewer’s identity and agreement. These are targets, not completed approvals.

Week 2: Target completed OpenAI, Supabase Auth and Razorpay integration; test sign-in, checkout and payment-backed access. Complete verse review and content permissions. Weeks 3–4: Test response quality, privacy, failure states, usage limits and cancellation; fix defects. Provider delays will delay paid launch.

Weeks 5–6: Run the campus trial only after content and AI checks pass. Fix issues and validate live payment readiness, cancellation and per-user usage limits.

Weeks 7–8: Invite trial users to the paid pilot and measure conversion, repeat use and operating cost. Delay paid release if content or safety checks fail.