cf-temp-dropper

name	cf-temp-dropper
description	Use when a user wants to share a local file through a short-lived Cloudflare Workers temporary deployment, especially via npx cf-temp-dropper.
version	0.1.0
author	kiyo-e
license	MIT
metadata	{"hermes":{"tags":["cloudflare","workers","file-sharing","temporary","npx","cli"],"homepage":"https://github.com/kiyo-e/cf-temp-dropper"}}

cf-temp-dropper

Use cf-temp-dropper when the user wants to turn a local file into a temporary public Cloudflare-hosted download/preview page.

The tool:

• splits one local file into temporary-account-safe Workers Static Assets chunks;
• generates a small Hono Worker and static UI;
• deploys with wrangler deploy --temporary;
• prints a short-lived workers.dev URL plus a Cloudflare claim URL;
• exposes /file for direct streaming/download with HEAD and byte-range support;
• provides a browser UI with media preview and checksum-verified save.

Quick command

npx --yes cf-temp-dropper@latest ./path/to/file

For reproducible tests, pin a version:

npx --yes cf-temp-dropper@0.1.0 ./path/to/file

Recommended agent workflow

1. Confirm the file exists before invoking the tool.

   test -f ./path/to/file && stat -c '%n %s bytes' ./path/to/file
   

2. Deploy through npx unless the user specifically asks to use a local checkout.

   npx --yes cf-temp-dropper@latest ./path/to/file --yes --keep
   

3. Capture both URLs from stdout:

   • public page URL: https://...workers.dev
   • claim URL: https://dash.cloudflare.com/claim-preview?...

4. Verify the public page and manifest.

   curl -sSIL https://<worker>.workers.dev/ | sed -n '1,12p'
   curl -sS https://<worker>.workers.dev/manifest.json
   

5. For direct file download, use /file.

   curl -L -o <filename> https://<worker>.workers.dev/file
   

   Resume interrupted downloads:

   curl -L -C - -o <filename> https://<worker>.workers.dev/file
   

6. Report the result succinctly: page URL, claim URL, file name, size, SHA-256, and any verification performed.

Dry run / local inspection

Use --no-deploy when the user wants to inspect generated files without uploading:

npx --yes cf-temp-dropper@latest ./path/to/file --no-deploy --out ./drop-build --yes
cd ./drop-build
npm install
npm run dev

Generated project layout:

package.json
wrangler.jsonc
src/index.ts
public/index.html
public/styles.css
public/app.js
public/manifest.json
public/chunks/part-00000.bin

Limits and expectations

• Temporary deployments are short-lived unless claimed in Cloudflare.
• The public URL is accessible to anyone with the link.
• Do not use this for secrets, credentials, regulated private data, or anything that must remain private.
• Default chunk size is 4.75 MiB for temporary-account safety.
• The tool refuses deployments that would exceed the 1,000 generated static-file limit.
• Browser verified-save uses IndexedDB and final Blob assembly, so very large files may hit browser storage or memory limits.
• /file is Content-Disposition: inline for media preview. curl -OJ https://.../file may save as file; prefer curl -L -o <filename> https://.../file.

Useful examples

Deploy an audio file and verify media headers:

npx --yes cf-temp-dropper@latest ./recording.mp3 --yes --keep
curl -sSIL https://<worker>.workers.dev/file | sed -n '1,20p'

Expected useful headers:

content-type: audio/mpeg
content-length: <bytes>
accept-ranges: bytes
content-disposition: inline; filename*=UTF-8''recording.mp3

Test a range request:

curl -sS -D headers.txt -H 'Range: bytes=0-99' -o segment.bin \
  https://<worker>.workers.dev/file
sed -n '1,20p' headers.txt
wc -c segment.bin

A good response is HTTP 206 with content-range and exactly 100 bytes for bytes=0-99.

Common pitfalls

• No URL appears: the command probably used --no-deploy; run without it for a live temporary URL.
• Cloudflare deploy uses the wrong account: the CLI intentionally clears common Cloudflare credential env vars before wrangler deploy --temporary, but local shell customization can still surprise you. Prefer the default npx path.
• Temporary URL disappears: expected for unclaimed temporary deployments; use the claim URL if persistence is needed.
• curl -OJ saves as file: use -o original-name.ext because /file is optimized for browser preview.
• Large file fails in browser save: preview may still work through byte ranges, but verified save can hit browser memory/storage limits.

Verification checklist

Before telling the user it worked, verify at least:

• deploy command exited 0;
• public URL was printed;
• curl -I <url>/ returns HTTP 200;
• curl <url>/manifest.json returns the expected file name, size, and SHA-256;
• if media/direct download matters, curl -I <url>/file returns accept-ranges: bytes and a correct content type.