lurk
Self-hostable Reddit buyer-intent finder that has Jev judge every post and comment it reads, with a written reason and cost per lead.
Overview
lurk finds people on Reddit who are likely buyers for your product. It builds a product profile from your URL (pain, solution, buyers, subreddits and searches), then scans on a schedule per keyword and subreddit: a title-level triage, opening only promising threads, a product-agnostic reading of each to see whether anyone is asking for anything, then the comments on the best ones. Per the listing, Jev judges every post and comment a scan reads. Each lead gets a 0 to 100 score from fit, intent and thread liveliness, an intent stage, the matching phrase, a flag for sellers rather than buyers, a one-sentence reason and the cost of the data behind it.
It also finds Reddit threads Google ranks for your keywords, tracks competitors' mentions and sentiment, groups leads into pain themes, sends alerts by email, Slack, Discord or webhook, and exposes everything over a read-only API and MCP. It deliberately never posts, drafts replies, archives Reddit or silently changes your filters. Data comes from AnyAPI on a pay-per-request basis; self-hosting is free and unlimited, with a hosted version at lurk.so.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Is this title worth opening (triage)?(inferred) | — |
| Yes / no | Is anyone in this thread asking for something?(inferred) | — |
| Score | How well does this post or comment fit the product, and how strong is the buying intent?(inferred) | — |
| Yes / no | Is this person selling rather than buying?(inferred) | — |
How it works
- Build a product profile from your URL
- Scan keywords and subreddits on a schedule
- Triage titles, then read promising threads and comments
- Score leads with a reason, stage and cost
- Alert and expose results via API and MCP
Strengths
- Cost scales with quality, not volume
- Reason and cost shown per lead
- No automated posting or reply drafting
- Self-hostable for free
Caveats
- Depends on AnyAPI for data
- Jev's exact questions not detailed in the README excerpt
- Lead scores are model judgments
Getting started
Follow the five-minute self-host section of the README, or use https://lurk.so.At a glance
- Type
- Project
- Status
- production
- Language
- TypeScript
- Runtime
- web app with scheduled scans, API and MCP
- Has code
- yes
Use cases and integrations
How curated lists describe it
Self-hostable Reddit buyer-intent finder that uses Jev to judge every post and comment a scan reads.
MrJev/awesome-jev
Frequently asked questions
Does lurk post on Reddit for me?
No, it never posts, comments or drafts replies.
What does it cost?
Self-hosting is free; Reddit data is paid per request through AnyAPI.
Can an agent use it?
Yes, via a read-only API and MCP.
Related projects
- jev-seo★ 610
Live SEO audit from one homepage URL: crawl, 52 rules, Core Web Vitals and Jev page judgments, reported as PDF, XLSX and Markdown.
PluginbetaPython Scores HeyReach LinkedIn leads and first-message variants with Jev, picks a message per lead and writes draft campaigns back.
ProjectbetaPythonMarketing and voice-agent platform with an experimental TypeSafe module that picks specialisation labels for agent roles.
Projectprototype- Jev Search★ 534
Plain-language web search where a decision model like Jev picks query, sources and time range, then ranks Search1API results.
AppproductionTypeScript Site audit for SEO and AI-search visibility: a deterministic crawl, narrow Jev Choice, Score and Noul questions, and a live dashboard.
AppbetaTypeScriptChrome extension that grades readable page sections on ten rubric scores with Jev via Vercel AI Gateway and gives the page an A-E grade.
ProjectbetaTypeScript