Skip to content

Latest commit

 

History

History
67 lines (45 loc) · 3.68 KB

PRD.md

File metadata and controls

67 lines (45 loc) · 3.68 KB

Product Requirements (PRD)

One-liner

LeadParrot finds public online conversations where people are already asking for help, recommendations, alternatives, services, or products — scores buyer intent, summarizes why each is a potential lead, and drafts a helpful reply you review and post yourself.

Core promise

Find people already asking for what you sell — and get a useful reply draft before your competitors see the thread.

Who it's for

SaaS founders, agencies, consultants, B2B service businesses, AI tools, and niche service businesses in US / EU / Canada / Australia.

What it explicitly is NOT

Not a generic social-listening platform, not a spam/automation tool, not a CRM, not LinkedIn/auto-DM/auto-comment automation, not a scraper of private/logged-in sources, not browser automation.

Product principles

  1. Safe & platform-conscious. Official APIs + public feeds only. Never post/DM for the user. Disclosure-first replies.
  2. Validation-first. Runs end-to-end with zero secrets (demo mode + mock AI). Narrow before broad.
  3. Lean & self-serve. Low support, simple metering, clear empty states.
  4. Working vertical slice over abstractions.

MVP feature set

  1. Landing page — hero, problem, solution, how-it-works, use cases, safety positioning, pricing, FAQ.
  2. Auth & onboarding — Supabase Auth; collect business + ICP + keywords + tone + digest prefs.
  3. Database + RLS — organization-scoped schema, no cross-org leakage.
  4. Dashboard — totals, high-intent count, avg score, saved, replies copied, active projects, last run, recent leads, usage meters.
  5. Projects — product description, ICP, competitors, keywords, negative keywords, geography.
  6. Sources — manual, Reddit (official API), Hacker News (public), RSS, web-search placeholder; each with "Run scan".
  7. Scan + scoring — keyword pre-filter → AI scoring (Zod-validated) → weighted overall score → lead candidate.
  8. Reply drafting — helpful, disclosed, non-pushy drafts with safety notes.
  9. Lead inbox — filter by project/source/score/stage/status, search, sort.
  10. Lead detail — full analysis, reply draft, copy (no posting), status, notes.
  11. Daily digest — email via Resend, or in-app preview.
  12. Billing — Free/Starter/Pro/Agency; Stripe abstraction; "not configured" fallback.
  13. Usage limits — monthly posts scanned, reply drafts, project count.
  14. Admin — aggregate internal metrics, gated by ADMIN_EMAILS.

Scoring model

overall = relevance*0.35 + intent*0.30 + urgency*0.20 + fit*0.15 (each 0–100). Tiers: high ≥ 70, medium 40–69, low < 40. Lead stages: research, problem-aware, solution-aware, buying-intent, competitor-switching, not-a-lead.

Plans & limits

Plan Price Projects Posts/mo Reply drafts/mo
Free $0 1 20 10
Starter $19 3 500 100
Pro $49 10 3,000 1,000
Agency $99 25 10,000 5,000

Acceptance criteria

Sign up/login · create org · create project (ICP/keywords/competitors/negatives) · add manual source/post · run mock AI scoring · lead inbox · generate reply draft · copy reply manually · save/mark leads · dashboard metrics · digest preview · billing/pricing abstraction · usage limits · RLS protection · platform-safety docs · validation plan · core tests pass.

Out of scope (v1)

Scheduled background scans, live web-search integration, team invites/roles UI, white-label digest, CRM features, outbound email/DM automation.

Success metric

See VALIDATION_PLAN.md: ≥ 3 of 10 trial users agree to pay $19–$49/mo; pivot/kill if < 2 of 10 find value.