Your research, front-page ready.
You spent years on the study. Now you have one headline, one hook and three key messages to win over a journalist — or lose them. Headline Hunter walks you through the classic media-prep worksheet: what you found, so what it matters, and now what should happen.
Drop in your paper (PDF, DOCX or plain text) and AI drafts headline ideas, a “so-what” hook and your key messages — grounded in what the paper actually says. Pick your favourite headline, watch it typeset live on a mock front page, then export a one-page press-release PDF.
How it works
Import your paper
Drag in a PDF, DOCX or text file. The text is extracted inside your browser — nothing is uploaded anywhere. Works offline, too.
Let AI draft the brief
Get 3–5 headline ideas, a one-line hook, and your what / so-what / now-what messages — written for everyday readers and grounded in the paper, with tone and target-outlet controls.
Make it yours & export
Every field stays editable — AI suggests, you decide. Re-roll headlines, watch the live front-page preview, then export a press-release PDF or print the worksheet.
Your paper never leaves your browser. File parsing happens locally with bundled libraries. If you use AI drafting, only the extracted text goes directly from your browser to the provider you choose — and your API key is stored only in this browser, never on our servers (we don’t have any).
Choose your edition
Browser edition
Zero install. Open the Studio, paste your own API key (Claude, OpenAI-compatible, or a local Ollama) in Settings — or skip keys entirely and use Copy prompt with any chatbot you already have.
Hosted edition
Self-host with Docker and your users never need a key: the server holds one API key (via .env), with an optional access passcode and per-IP rate limiting built in.
docker compose up -d
Local edition
It’s just HTML — clone the repo and double-click studio.html, or run the tiny launcher to auto-fill keys from your environment.
git clone github.com/michael-borck/headline-hunter python3 serve.py