ideas you can hear
before they exist
a personal tool that listens to what people complain about online, and turns it into projects worth building
2,542 posts read · 29 ideas found · last listened 12 days ago
enter →the problem
you want to build something, but you don’t know what.
every idea you think of is either too large to finish, or something nobody actually needs.
meanwhile people describe their problems openly, every day, in public.
Annoying and alarming things about OpenCode
— hacker news, july 2026
how it works
reads public posts from hacker news and reddit
a cheap model discards anything that isn't a real complaint
a second pass pulls out the problem, who has it, and what a minimum version would look like
urgency is calculated in code from measurable signals
an example
Claude Code/Cowork sering gagal menjalankan sandbox di lingkungan Windows sehingga tidak bisa mengeksekusi kode secara langsung, membuat fitur utamanya tidak berjalan.
target user
Software engineer yang menggunakan Claude Code atau AI coding assistant di sistem operasi Windows
mvp scope
Daemon lokal ringan yang memantau file pemicu dan menjalankan skrip PowerShell/CMD secara aman serta mengembalikan outputnya.
effort · source
weekend · hacker news
score breakdown — 46/100
- engagement—16
- pain intensity—100
- willingness to pay—0
- recency—60
- feasibility—100
Pengguna Claude desktop (Pro/Max) harus terus-menerus mengklik ikon roda gigi untuk memeriksa sisa batas penggunaan (usage limits), yang mengganggu perencanaan sesi kerja mereka.
Developer dan profesional yang menggunakan Claude desktop app dengan langganan Pro atau Max. · 1 week · hacker news
Bisnis e-commerce yang bertumbuh cepat menghadapi penutupan akun sepihak dan penahanan dana oleh satu payment processor tanpa mekanisme fallback otomatis.
Pemilik bisnis e-commerce dan merchant digital berskala menengah dengan volume transaksi cepat meningkat. · 1 month · hacker news
this is real output, not a mockup
how urgency is scored
the score is calculated in code, not written by the model. the model only extracts variables — the weighting is deterministic and auditable.
each component is stored per idea, so any score can be traced back to its inputs.
why this exists
i kept opening a blank editor with no idea what to build.
the ideas i came up with alone were either too ambitious to finish on a weekend, or solved a problem only i had.
so i built something to listen for me.
what it doesn't do
it does not predict whether an idea will succeed.
it does not validate market demand — a loud complaint is not the same as a paying customer.
it does not replace talking to the people who have the problem.
where it listens
currently hacker news and reddit. github issues and product review sites are next.
public posts only. no personal data is stored — usernames are hashed before they touch the database.