Sales outreachCAN I VIBECODE
SubSignal?
BUILD FIRST
- A non-sensitive or read-only version of the core experience
- A manual review step wherever correctness or trust matters
- A narrow companion tool that removes a real piece of work
LEAVE FOR LATER
- Regulated decisions, payments, or irreversible actions
- Claims of correctness, compliance, or uptime that the MVP cannot support
- Sensitive integrations until security, support, and auditability are ready
RESEARCH NOTES
What this tool actually does.
SubSignal (subsignal.ai) is a Reddit monitoring and sales lead generation SaaS. It scans subreddits 24/7, uses AI to score posts by relevance and buying intent (0-100 match scores; intent filters: buying, researching, comparing), and delivers high-intent opportunities to the user's inbox. Features include brand monitoring (instant alerts for brand, product, or competitor mentions), real-time alerts, AI content detection (flags ChatGPT-generated reviews/testimonials), an opportunity finder, a user/account scanner that analyzes up to 1,200 posts to judge whether an account is genuine, a bot, or a shill, and Reddit DM automation that sends AI-personalized DMs crafted around the person's exact post. Pricing: Growth $29/month, Professional $129/month for teams and agencies. Marketing claims it saves 10+ hours per week versus manual prospecting.
RESEARCHED BUILD PROMPT
Give this to your coding agent.
Build a focused replacement for SubSignal, a Reddit monitoring and sales-lead tool. Workflow: a user registers, connects their product website/domain and Reddit account, defines tracked keywords (brand, product, competitor names, problem phrases), and the system continuously ingests Reddit posts/comments via Reddit's official API (Pushshift is deprecated; use Reddit's search/streaming endpoints). Each post is scored by an LLM (e.g., GPT-4o-mini) for relevance and buying intent (buying/researching/comparing) with a 0-100 match score; high-scoring posts land in an inbox with email/notification alerts. Users filter by intent, reply inline, and optionally send AI-personalized DMs drafted from the specific post content. Include a user scanner that fetches up to ~1,200 posts from an account and returns a bot/shill/genuine verdict. Tech approach: Next.js front end, Postgres with a queue (BullMQ/Redis) for ingestion, an LLM scoring pipeline, and Reddit OAuth. User-facing requirements: dashboard, inbox, filters, alerts, DM composer, account scanner, pricing tiers. Do not copy SubSignal's brand, copy, logo, or UI. Out of scope: multi-platform monitoring (Twitter, LLM responses), agency/team billing, mobile apps. Deliver a README covering setup, Reddit API rate limits, LLM cost estimates, and limitations (Reddit API restrictions and DM ToS compliance).