Find blog topic ideas based on real questions buyers are asking about competitors on Reddit and X.
Discover live conversations where your brand could reply, ranked by freshness and how answerable they are.
Detect competitor astroturfing, such as fake grassroots complaints or seeded keyword campaigns.
| juicer-io/skills | 787a68/hubproxy | 920linjerry-stack/capital-studio | |
|---|---|---|---|
| Stars | 16 | 16 | 16 |
| Language | — | Go | Python |
| Setup difficulty | easy | moderate | easy |
| Complexity | 1/5 | 3/5 | 3/5 |
| Audience | pm founder | ops devops | researcher |
Figures from each repo's GitHub metadata at analysis time.
First run signs you up for a free API key by email, no dashboard setup needed.
This repository is a collection of skills for AI coding agents like Claude Code, aimed at doing real marketing research using live social media data. It connects to the Juicer Data API, which gives one unified way to pull information from Reddit, X, Instagram, TikTok, Facebook, YouTube, LinkedIn, and other platforms without needing separate logins or scraping setups for each one. When you first use a skill, it signs you up for a free API key using just your email address, so there is no dashboard to configure ahead of time. Three skills are included. Topic scout searches Reddit and X for the actual questions buyers are asking about your competitors, along with live complaints and outage reports, and it can spot when a company is faking grassroots buzz for its own product. Mention scout finds current conversations where your brand could genuinely join in, such as fresh requests for tool recommendations, competitor complaints, or unanswered mentions of your own brand, and ranks them by how fresh and answerable they are. The third skill, simply called juicer, is not a workflow but a reference that teaches an agent how to use the Juicer API directly, covering feeds, moderation, embed codes, webhooks, and cross platform lookups, for people who want to build their own tool on top of it. Installation works through the Claude Code plugin system with a few commands, or by copying a skill's folder directly into any compatible agent's skills directory, since each skill is just a folder of instructions with no extra runtime beyond common command line tools. Every finding these skills produce includes a source link, and the outputs are meant to guide a human's decision rather than publish anything automatically. This is aimed at marketers and founders who want fast, evidence backed research instead of manually searching social platforms themselves.
A set of Claude Code compatible skills that use live social media data to find real marketing conversations, blog topic ideas, and brand mentions worth responding to.
MIT license: use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly pm founder.
This repo across BitVibe Labs
double-check against the repo, no cap.