A configurable in-memory Dat-to-HTTP gateway, so you can visit Dat archives from your browser.
If you want a browser that can visit Dat archives, check out Beaker.
To get the dat-gateway command for running your own gateway, use npm:
npm i -g dat-gateway
If you have npx installed, it's even shorter:
npx dat-gateway
- http://gateway.mauve.moe:3000/ (Hosted by @RangerMauve)
- https://pamphlets.me/ (Hosted by @brechtcs)
- https://dat.bovid.space/ (Original gateway from @garbados)
- https://dat.hypersource.club (Hosted by @jwerle)
You can run dat-gateway to start a gateway server that listens on port 3000. You can also configure it! You can print usage information with dat-gateway -h:
$ dat-gateway -h
dat-gateway
Options:
--version Show version number [boolean]
--config Path to JSON config file
--host, -l Host or ip for the gateway to listen on. [default: "0.0.0.0"]
--port, -p Port for the gateway to listen on. [default: 3000]
--dat-port, -P Port for Dat to listen on. Defaults to Dat's internal
defaults. [default: null]
--dir, -d Directory to use as a cache.
[string] [default: "~/.cache/dat-gateway"]
--max, -m Maximum number of archives allowed in the cache. [default: 20]
--period Number of milliseconds between cleaning the cache of expired
archives. [default: 60000]
--ttl, -t Number of milliseconds before archives expire.
[default: 600000]
--redirect, -r Whether to use subdomain redirects [default: false]
--loopback, -L What hostname to use when serving locally.
[default: "dat.localhost"]
-h, --help Show help [boolean]
You can visit Dat archives through the gateway using a route like this:
http://localhost:3000/{datKey}/{path...}
For example:
http://localhost:3000/40a7f6b6147ae695bcbcff432f684c7bb5291ea339c28c1755896cdeb80bd2f9/assets/img/beaker-0.7.gif
The gateway will even resolve URLs using Dat-DNS:
http://localhost:3000/garbados.hashbase.io/icons/favicon.ico
The gateway will peer archives until they expire from the cache, at which point it proactively halts them and deletes them from disk.
The gateway also supports replicating a hyperdrive instance using websockets
const websocket = require('websocket-stream')
const hyperdrive = require('hyperdrive')
const key = 'c33bc8d7c32a6e905905efdbf21efea9ff23b00d1c3ee9aea80092eaba6c4957'
const url = `ws://localhost:3000/${key}`
const archive = hyperdrive('./somewhere', key)
archive.once('ready', () => {
const socket = websocket(url)
// Replicate through the socket
socket.pipe(archive.replicate()).pipe(socket)
})By default dat-gateway will serve all dats from the same origin. This means that dats using absolute URLs (starting with /) will be broken.
This also means that all dats will share the same localStorage and indexedDB instances which can cause security issues.
In order to resolve these issues, you can use the --redirect flag in conjunction with the host parameter to have each dat served on a subdomain.
For example, http://{host}:{port}/{datkey}/index.html will be redirected to http://{datkey32}.{host}:{port}/index.html which will serve the file from localhost, but at a different domain, ensuring the browser isolates all the contents from each other.
Please note that due to limitations in how URLs work, the dat key will be converted to it's base32 representation instead of hexadecimal using this library
Running a gateway locally for personal use is a great idea, but by default dat-gateway uses dat.localhost as its hostname when serving to the local machine. Firefox does not support '*.localhost' domains and so this behavior breaks the gateway for Firefox users.
To fix this, use the -L, --loopback flag to specify localhost as the loopback hostname, like so:
$ dat-gateway -L localhost
This will cause dat-gateway to use localhost as its domain name, which Firefox supports just fine.
All contributions are welcome: bug reports, feature requests, "why doesn't this work" questions, patches for fixes and features, etc. For all of the above, file an issue or submit a pull request.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent