calcLLM

Deploy calcLLM to production on Vercel with all infrastructure provisioned — production database, Stripe live keys, DNS, and GitHub OAuth production app. The product should be accessible at calcllm.com.

Prepare for public launch — fix rough edges discovered during deployment, polish error states, improve landing page conversion, and begin beta outreach. The product should be "Show HN ready."

Public launch — Show HN, sustained Twitter/X engagement, and first 10 paying customers. Begin the 30/60/90 day traction plan from research/gtm.md.

Broaden self-serve authentication beyond GitHub without breaking existing users, saved estimates, or team sharing. Google is the first post-launch auth expansion because it reduces signup friction for technical founders while preserving the current PLG motion.

Enable passwordless sign-in for non-GitHub/non-Google users and remove the GitHub-username constraint from collaboration. This phase must ship with abuse protections from day 1.

Support larger teams and enterprise buyers with domain-based org identity and enforceable SSO, without compromising the self-serve auth flows that serve the primary ICP.

Build the mathematical core of calcLLM — the calculation engine that computes realistic LLM costs for 6 workload templates. This is the foundation everything else builds on. Starts with test infrastructure since none exists.

Build the primary conversion driver — the workload estimator that creates the "aha moment." Users select a use case, configure parameters with sliders, and see their naive vs realistic costs update in real-time.

Make the product launch-ready — mobile responsiveness, shareable URL state, SEO optimization for the pricing table, and smooth CTA flow from table to estimator.

Add the backend foundation for persistent features — GitHub OAuth authentication and Neon Postgres database with Drizzle ORM schema for users and estimates.

Build the freemium value — save estimates, version them, share via permalink, and collaborate with team members. This is what converts free users to registered users.

Enable the paid tier — users connect their LLM provider API keys for live cost tracking, and subscribe via Stripe for access to tracking features.

Deliver the paid tier's retention value — actionable optimization recommendations, scenario modeling, and reporting tools that justify the subscription cost.

Integrate PostHog Cloud to unblock all metric tracking from research/metrics.md. Every visitor from day 1 of deployment should be tracked. This must ship before Phase 9 (deploy) so there's no data gap.