WITH CLAUDE logo WITH CLAUDE Search

LaunchJury

A command-line agent that reads a repository, drafts a complete product launch and tests the drafts against a jury of seven audience personas before you post.

Built at Fable 5.1 Build Day · BhopalAgentFunctional · self-reported

The problem

Indie developers and small teams finish building side projects and then never launch them properly. Writing a Show HN post, a Product Hunt tagline, a demo video script, and figuring out which communities to post in feels like a second project, so most projects get one rushed tweet and die quietly. The people who feel this are solo builders who are good at shipping code but have never written a launch post, do not know how Hacker News or Reddit will react, and cannot afford a marketer. The result is good work that never reaches the audience it was built for.

How it works

LaunchJury is a CLI agent built on the Claude Agent SDK that turns a repo into a complete, pre-tested launch. You point it at a repository. It reads the code and README to understand what the product actually does, then researches how similar products were received on Hacker News, Reddit, Product Hunt and Indie Hackers, pulling real comments with URLs. It scores the positioning against eight launch principles (one-liner clarity, differentiation vs. found competitors, objection coverage, proof, and so on) with evidence and a concrete fix for each. It then drafts every asset: a Show HN post and first comment, Product Hunt tagline and maker comment, an X thread, per-subreddit Reddit posts, a LinkedIn post, a scene-by-scene demo video script with a shot list, and a 60-90 second launch video script. The drafts go before a jury of seven specialist personas (Hacker News veteran, Product Hunt coach, Reddit moderator, X growth marketer, dev-tools copywriter, demo director, skeptical target user). Each scores the assets, predicts the literal top comment the audience would leave, and lists must-fix edits. A deterministic foreman aggregates the weighted scores and sends drafts back for revision until they pass. Finally it produces a distribution plan (which channels, what to post where, best day and time, who to tag, what rules to respect, a day-by-day sequence) and a single self-contained HTML report with everything. It solves the problem by doing the research, validation, drafting, and audience-testing a marketer would do, before you post, so the launch hits the right audience in the right place with copy that has already survived the critics.

How Claude was used

In the team’s own words, from their submission.

LaunchJury is a CLI agent built on the Claude Agent SDK that turns a repo into a complete, pre-tested launch.

Team

beacon

Individual builders are not listed yet — names are published only with their permission.

← Back to the Project Directory