Lynqd
Overview
Lynqd is a browser-based outreach engine built for freelancers and consultants who do volume client outreach without wanting to sound like it. You define reusable presets — identity, tone, strengths, prompt templates, output language — then for each client you add what you know about them, optionally select which services you're pitching and whether to pull in Knowledge Base context, and the app calls AI through its own API routes to produce a personalized email, a proposal PDF, or both. Outputs are fully editable in-app, with per-section AI regeneration and a PDF template switcher. A separate Social tab handles X-format content generation — single tweets and threads — using the same services and Knowledge Base as context but completely isolated from the outreach flow. Everything runs locally in the browser with no backend, no accounts, and no data leaving your machine. Built on Next.js 15, Dexie, and @react-pdf/renderer with a full import/export backup system.
Features
- Built a full preset system where each preset carries its own identity, tone, prompt templates, default services, and output language — editing one preset never affects others.
- Engineered a two-mode output system (Email only / Email + PDF) with separate AI calls per format, purpose-built prompts for each, and a PDF template switcher that requires no additional API call.
- Implemented a global Services library and Knowledge Base (manual text + .txt/.pdf/.docx ingestion) that inject structured context into every generation with offer-depth behavior scaling by selected service count.
- Built a per-section AI regeneration system on both email and PDF outputs — select a section, give optional instructions, replace only that piece without touching the rest.
- Designed a fully isolated Social Media Content Engine for X — own storage, own API route, own profile and post-type system with seeded defaults — removable without touching the outreach flow.
- Local-first architecture with no backend, no accounts, and no hosted data — all app data in IndexedDB via Dexie, config in localStorage, full backup and restore via a .lynqd JSON file.
- PDF generation runs entirely client-side via @react-pdf/renderer with three swappable proposal templates that inherit global branding (logo, colors, contact block).
- Built on Next.js 15 App Router with Dexie, TanStack Query, react-hook-form, zod, and shadcn/ui — provider-agnostic AI layer supporting Chutes, Gemini, OpenRouter, and Ollama.