Try it: stack-converter.onrender.com
Paste a public Substack article URL. Get the article back as ready-to-paste LinkedIn Article and X Article formats, with the images, captions and embedded video carried over, and four AI-generated opening hooks per platform that are grounded strictly in the article's own text.
Built for people who write on Substack and republish on LinkedIn and X without wanting to reformat by hand or let an AI rewrite their voice.
- Fetches the article from Substack's public API (any author, any publication) and parses the full post: headings, paragraphs, lists, quotes, links, inline images and
<figcaption>captions. - Rebuilds it for each platform. LinkedIn Articles and X Articles accept different formatting; the converter produces one version for each, copied to the clipboard with a single button.
- Carries media over. Images are fetched at full resolution; embedded Substack video is served back so it can be re-attached.
- Writes hooks, not rewrites. Four opening hooks per platform (different angles: contrarian, practical, story, takeaway), each under 240 characters, generated from the article text only. Clichés and marketing phrases are banned in the prompt. Any hook can be refreshed individually.
- Falls back gracefully. If no model key is set or the model call fails, hooks are synthesised from the article's own first sentences so the tool still works.
The interface is a single page: one URL field, one Convert button, the article preview, and copy buttons for title, LinkedIn and X.
Requires Python 3.12+. No third-party Python packages: the server uses only the standard library.
git clone https://github.com/zahedsaifi94/stack-converter.git
cd stack-converter
export GEMINI_API_KEY=your_key # optional; see "AI hooks" below
./start.sh # serves on http://127.0.0.1:8765Open the URL it prints, paste a Substack link, click Convert.
The hosted instance above runs on Render's free plan, so it may take a few seconds to wake up. To run your own copy: the repo includes a Dockerfile and a render.yaml, so you can deploy it to Render in a few clicks and set GEMINI_API_KEY as an environment variable in the Render dashboard. The service listens on $PORT (default 8080 in the container) and exposes /healthz and /ping for uptime monitors, so a free-tier instance can be kept awake with a cron ping.
It will run anywhere that can run a Docker image or a plain python3 server.py.
Hooks are generated with Google's Gemini API by default. Configuration, all via environment variables:
| Variable | Purpose |
|---|---|
GEMINI_API_KEY (or GOOGLE_API_KEY, API_KEY) |
Gemini API key from Google AI Studio |
GEMINI_MODEL |
Override the model; otherwise the server tries current Gemini 3 models in order |
GROQ_API_KEY / OPENAI_API_KEY |
Optional. If set (or if a gsk_/sk- key is pasted into GEMINI_API_KEY), the server uses that provider instead |
APP_CSRF_SECRET |
Optional. Fixed secret for session tokens; otherwise one is generated and stored in .csrf_secret |
Without any key the tool still converts articles; it just uses the text-derived fallback hooks.
server.py the whole backend: Substack fetch, parsing, platform formatting, hook generation, HTTP server
public/ the single-page frontend (index.html, app.js, styles.css)
media_vault/ runtime cache for fetched media (ignored by git)
Dockerfile container image
render.yaml one-click Render deployment
start.sh local launcher
- Works with public Substack posts only; paywalled content behind a subscription is not fetched.
- The converter is deliberately conservative: it reformats, it does not rewrite. The only generated text is the hooks, and they are constrained to the article's content.
- Fetched articles are cached in memory for the life of the process.
MIT. See LICENSE.