XG 3.3.3.2 — Mono/Docker Compatibility Fork

This fork revives XG (XDCC Grabscher) 3.3.0.0 for modern Mono and Docker environments.

XG was originally created by Lars Formella. This fork preserves the original project, licensing, and attribution while applying compatibility and maintenance fixes needed to run XG on current systems.

This compatibility release is based on XG 3.3.0.0 and reports itself as XG 3.3.3.2.

What's changed

3.3.3.2

• A bot sitting in several of your channels, also on different networks, has its packets stored once instead of once per channel; duplicates already stored are removed at startup and every 20 minutes, keeping the copy of an online bot and never touching a packet that is queued or downloading
• The database is saved at most every 2 seconds instead of after every change, which made each new packet of a bot with a large pack list (100,000+) cost a quarter of a second or more

3.3.3.1

• Much less memory and CPU with large pack lists: a bot announcement no longer copies the bot's whole pack list or rebuilds the IRC parser's patterns, so XG's memory stays flat instead of climbing by gigabytes between garbage collections

3.3.3.0

• Prowlarr, Sonarr and Radarr integration: XG works as a Newznab indexer and a SABnzbd-compatible download client (see below)
• Passive (reverse) DCC support with forwarded ports, ready for VPN setups (see Passive DCC)
• Optional download folder per category (dl/tv, dl/movies)
• More reliable downloads:
  • refused or silent DCC connections are retried instead of giving up on the first try
  • stale offers are cancelled, and offers are matched to packets by file name
  • a broken transfer is never finished as complete, and parts are resumed, also from another bot
  • finished files never replace existing ones (saved as name (1).ext)
  • the next packet from a bot, or a bot waiting for a free slot, is requested within seconds instead of up to 4 minutes
  • bots that never answer or send unusable offers end the request, so the *Arr can try another release
• Stability and security:
  • a malformed DCC line from anyone in a channel no longer crashes XG
  • IRC messages are never handled twice, and queues shared between threads are locked
  • pack lists are only downloaded when XG asked for them
• Packet names keep non-ASCII letters (umlauts, accents), and search ignores accents and apostrophes
• Fixed the crash on the very first start while creating xgsnapshots.db
• Fixed /api/1.0 requests hanging and wrong responses
• Download folders changed in the web interface work without a restart
• A lab with a fake IRC network and scripted bots for testing XG without real IRC (tools/lab)

3.3.2.0 and earlier

• Builds and runs with Mono 6.12
• Added a reproducible multi-stage Docker build
• Fixed Mono build detection for current Mono/Roslyn toolchains
• Removed the obsolete Jabber plugin dependency
• Uses the Debian/Mono-compatible db4o assembly at runtime
• Forces SignalR to use long polling instead of unsupported WebSockets
• Automatically reconnects the web UI after SignalR disconnects
• Fixed packet searches hanging when the first search was entered manually
• Removed the obsolete remote-settings loader
• Removed the defunct XG cloud-search configuration and donation/server-status message
• Added native IRC TLS support, with automatic TLS on port 6697 and explicit ircs:// support for custom TLS ports
• Validates IRC server certificates when using TLS
• Fixed NickServ authentication and reconnect handling, including protected-channel retries after identification
• Prevented NickServ credentials from being sent as an IRC server password and redacted authentication credentials from logs
• Fixed the IRC password field in the web interface

Docker

Prebuilt image

Every release is published to the GitHub Container Registry:

ghcr.io/captainepix/xdcc-grabscher:latest
ghcr.io/captainepix/xdcc-grabscher:3.3.3.2-mono2026

latest follows the newest release; the versioned tag stays fixed. Use either one instead of xdcc-grabscher:3.3.3.2 in the examples below if you do not want to build the image yourself.

On unRAID, install XDCC-Grabscher from Community Applications. The template already contains the ports, paths and the optional settings described below.

Build

From the repository root:

docker build -t xdcc-grabscher:3.3.3.2 .

Run

docker run -d \
  --name xdcc-grabscher \
  --restart unless-stopped \
  -p 5556:5556 \
  -v /path/to/config:/config \
  -v /path/to/downloads:/config/.config/XG/dl \
  xdcc-grabscher:3.3.3.2

Then open:

http://YOUR-SERVER:5556

Change the default XG web password (xgisgreat) after your first login.

XG deliberately refuses to run as root. The Docker image therefore runs XG as UID/GID 99:100.

Make sure the directories mounted at /config and /config/.config/XG/dl are writable by that user. For example:

mkdir -p config downloads
chown -R 99:100 config downloads

The exact host paths are up to you; the paths above are only examples.

Docker Compose

The repository includes a docker-compose.yml using local config and downloads directories.

Before starting it for the first time:

mkdir -p config downloads
chown -R 99:100 config downloads
docker compose up -d --build

Persistent data

XG stores its configuration and application data underneath:

/config

Downloads are written by default to:

/config/.config/XG/dl

Both should normally be backed by persistent Docker bind mounts or volumes.

Passive DCC

Most bots open a port and XG connects to them. Some bots only offer passive (reverse) DCC: XG has to open a port, tell the bot its public address, and the bot connects to XG. This needs a few ports that are reachable from the internet.

1. Pick a few ports, e.g. 50000-50004. Each port carries one passive transfer at a time.
2. Forward them on your router (TCP) to the machine running XG.
3. Publish them on the container with the same numbers, e.g. -p 50000-50004:50000-50004 (on unRAID add them as ports, or put that into Extra Parameters).
4. Tell XG with environment variables:

docker run -d \
  --name xdcc-grabscher \
  -p 5556:5556 -p 50000-50004:50000-50004 \
  -e XG_PASSIVE_DCC_PORTS=50000-50004 \
  -v /path/to/config:/config \
  -v /path/to/downloads:/config/.config/XG/dl \
  xdcc-grabscher:3.3.3.2

Behind a VPN (e.g. XG sharing the network of a gluetun container): the ports have to be forwarded by the VPN provider, and not every provider offers that. Because XG looks up its address through its own connection, it finds the VPN's address by itself. If the provider forwards a fixed port, set it in XG_PASSIVE_DCC_PORTS; if the forwarded port changes on every connection, share gluetun's port file with XG and point XG_PASSIVE_DCC_PORTS_FILE at it.

Only the address a bot announced may connect to a passive port; the port is closed again after the transfer or after 90 seconds without a connection. Without XG_PASSIVE_DCC_PORTS, passive offers are rejected as before, and bots which sent one are left out of the Prowlarr/Sonarr/Radarr search results for 7 days.

Prowlarr / Sonarr / Radarr integration

XG can act as an indexer for Prowlarr and as a download client for Sonarr, Radarr and Prowlarr. Nothing has to be patched or installed next to XG:

Sonarr/Radarr → Prowlarr → XG Newznab API → NZB file → XG SABnzbd API → XDCC download → Sonarr/Radarr import

To do that, XG offers two small compatibility APIs on its normal web port:

Please note:

• XG is not a Usenet client. No Usenet server is ever contacted.
• The NZB files XG hands out are only an internal envelope that carries the XG packet through Prowlarr/Sonarr/Radarr and back to XG. They are useless to a real Usenet downloader.
• The SABnzbd API exists only so the existing *Arr SABnzbd client can submit and monitor XDCC downloads.
• Search is text only. IMDb/TVDB/TMDB id searches are not supported, because XDCC packets only have a file name.

1. Create an API key in XG

In the XG web interface, open the settings menu (gear icon), choose Api Keys and add a key with a name (for example arr). New keys start disabled: click the icon at the left of the key's row to enable it. The long value in the Api Key column is what Prowlarr, Sonarr and Radarr need.

If Prowlarr or the *Arr test reports Incorrect user credentials or API Key Incorrect, the key is usually still disabled.

You can use one key for everything or one key per application.

2. Prowlarr: add XG as indexer

Indexers → Add Indexer → Generic Newznab

The Prowlarr test asks XG for its newest packets, so at least one bot must be online with packets when you press Test.

Prowlarr then syncs XG to Sonarr (TV) and Radarr (Movies) like any other indexer.

By default only packets of bots that are currently online are returned, because offline bots can not deliver. To include offline bots anyway, set Additional Parameters to &offline=1.

3. Sonarr and Radarr: add XG as download client

Settings → Download Clients → + → SABnzbd

Prowlarr only provides the indexer; each *Arr application still needs XG configured as its download client.

If you want to grab directly from a Prowlarr search, add the same SABnzbd download client in Prowlarr too (its default category prowlarr exists in XG).

If you also use a real SABnzbd (or any other Usenet client), Sonarr and Radarr may send XG results to it, because both look like Usenet to them. Tie the XG indexer to the XG download client: in Sonarr/Radarr open Settings → Indexers → XG, click Show Advanced, and set Download Client to the XG client. Do the same for your Usenet indexers and their client. If a Prowlarr sync ever resets it, set it again.

4. Remote Path Mapping

XG reports finished downloads with the path XG itself sees, by default inside its download folder:

/config/.config/XG/dl/Some.Show.S02E05.720p.mkv

XG can not know how that folder is mounted into your Sonarr or Radarr container. If the paths differ, add a Remote Path Mapping in Sonarr/Radarr (Settings → Download Clients → Remote Path Mappings).

Example on unRAID, where the same share is mounted differently into both containers:

Remote Path Mapping in Sonarr:

If you changed the download folder in the XG settings, XG reports that folder instead.

5. Optional: a folder per category

By default every download lands directly in XG's download folder. With the environment variable XG_CATEGORY_FOLDERS, downloads grabbed by an *Arr go into a subfolder named after the category it sent (tv from Sonarr, movies from Radarr, as set in their download client):

Packets you start yourself in the XG web interface have no category and stay in the download folder, so other tools (like FileBot) can tell them apart from *Arr downloads. Category names only become folders if they are plain names (letters, digits, space, ., _, -). An existing Remote Path Mapping for the download folder also covers the subfolders.

How it behaves

Searching

• Text queries are matched against packet file names the same way the XG search does. Dots, dashes and underscores count as word separators, and -word excludes a word.
• Sonarr episode searches become S02E05. Season searches match every packet whose name contains a token starting with S02. Daily shows are matched by their date.
• XG does not know whether a packet is a movie or an episode. The category comes from the request: movie searches are returned as Movies, TV searches as TV, other searches in the requested category or as Other (8000).
• The publish date is when the packet started to offer its file (when XG first saw it, or when the bot put a different file under that pack number). Bots announce their packets over and over, so the announcement time would make old packets look new. The feed without a search term (RSS) is sorted the same way.
• Bots reuse pack numbers. If a packet offers a different file by the time it is grabbed, the grab is refused, so the wrong file is never downloaded.

Downloading

• A grab enables the packet in XG, exactly like clicking it in the web interface.
• The *Arr queue shows the job as queued while XG waits for the bot, and as downloading with progress and speed during the transfer.
• When XG has moved the finished file into its download folder, the job moves to history as completed, including the file path, and the *Arr imports it.
• Some bots only accept the file transfer on part of the ports they offer. If the transfer connection fails, XG cancels the bot's offer and asks again, up to 3 times, before giving up.
• If XG stops the download (transfer connection failed 3 times, bot offline, request denied, pack no longer valid, or disabled by you in XG), the job is marked failed with the reason and the last bot message. The *Arr can then search for another release or retry.
• A bot that does not answer at all fails the job after 15 minutes, so the *Arr can try another release. Bots that answer, for example with a queue position, are waited for as long as it takes. XG_COMPAT_SILENT_BOT_SECONDS changes the limit.
• Removing a completed item from the *Arr history only forgets the job. The downloaded file is only deleted when the *Arr explicitly asks to remove the data, and only if it is still that job's file inside XG's download folder.
• Jobs are stored in /config/.config/XG/arr-jobs.json and survive restarts.

Troubleshooting

The XG log (docker logs xdcc-grabscher) shows what happened to each grab: the request to the bot, its offer, and why a transfer failed.

Known limitations

• Text search only; no IMDb/TVDB/TMDB id lookups, no anime absolute episode numbers, and names like 2x05 are not found by episode searches.
• Removing a downloading item from the *Arr queue stops the XDCC transfer, and XG always deletes the partial file, even if the *Arr was asked to keep data.
• XG has no pause or priorities. SABnzbd priorities are accepted but ignored.
• Bots that only offer passive (reverse) DCC need forwarded ports, see Passive DCC.
• Categories only get separate folders when XG_CATEGORY_FOLDERS is set (see above).
• Grabbing a packet that is already being downloaded for an *Arr returns the existing job instead of a second one.
• The reported SABnzbd version is a fixed compatibility value.
• API keys are passed in the URL, as Newznab and SABnzbd clients expect. Keep XG on your LAN (see below) and do not share these URLs.

Security warning

XG is an old application and still depends on an old web stack, including legacy versions of Nancy, SignalR, jQuery, Bootstrap, and other libraries.

Do not expose the XG web interface directly to the public Internet.

For normal use, keep it accessible only on a trusted LAN. If remote access is required, place it behind an appropriately secured reverse proxy with authentication and TLS.

Modernizing the dependency stack is outside the scope of this initial compatibility release.

Known issues

SignalR I/O warnings under Mono

The container log may occasionally contain warnings similar to:

SignalR exception thrown by Task: System.AggregateException:
One or more errors occurred. (I/O error occurred.)

These appear to be associated with SignalR long-poll connection turnover under Mono. During testing they have not interrupted the web interface, searches, IRC connectivity, or XDCC downloads.

Truncated API answers under very heavy load

At a few hundred requests per second the embedded web server (Nowin) now and then cuts an answer off after 8 KB, about once in 20,000 requests in the lab's torture test. Prowlarr, Sonarr and Radarr send a few requests per minute and simply ask again on their next poll.

WebSockets

The legacy SignalR/Nowin stack used by XG does not provide working WebSocket support under the current Mono environment. This fork explicitly uses SignalR long polling instead.

Release naming

Application version:

3.3.3.2

Compatibility release/tag:

v3.3.3.2-mono2026

Original project

Everything below this point is the original XG project documentation.

XG

XG, called __X__dcc __G__rabscher, is a XDCC download manager. Grabscher is the german word for grabber :-)

What makes it special?

XG is just a command line app which connects to one or multiple IRC networks and handles the whole network communication. The IRC servers, channels, bots and packets are presented within a nice and stylish web frontend. There you can search and download packets.

You can run XG on every machine that supports C# / Mono - even root servers without x(org), or an old weak pc running linux without a monitor - and control your downloads with your browser from everywhere. You don't have to keep a big PC running, but just a small download box which handles all the IRC stuff.

How do i use it?

Run the program and point your browser to 127.0.0.1:5556. The default password is xgisgreat. If you already added some servers and channels, it will take some time untill the Webfrontend is up and running. This is due to the build in SQLite database which is not really performant and takes some time to load the saved objects.

Password Dialog

At first: change the settings

You can do this directly in the web frontend. Just click on the Config link in the options menu.

Options

This is a small explanation to help you set the correct options. If you don't want to use a special feature, just disable it.

Note: The Elastic Search configuration is not available in the webfrontend anymore and can be changed by editing the config file manually.

Settings part 1

The web server password is filled with xgisgreat and the port ist 5556. The IRC passport and email can be left blank and are just needed if you want use nickserv.

Settings part 2

Filehandlers

If a packet is downloaded you can run several commands. If the regex of a file handler matches the file name, the process is started. A process is defined by a command, arguments and the next process. The next process can be left empty and only is called if the current one is successfully executed.

Settings part 3

The following handler matches all rar / zip archives. It will create a separate folder, extract the archive into it and removes the archive. Every process is executed only, if the previous one was successfully. Because of this, the handler won't delete the archive if he could not extract it.

Settings part 4

You can add as many file handlers as you want. They are also stored in the settings file.

Arguments

You can use different placeholders in your arguments:

• %PATH% = full path of the file, like /the/full/path/to/file_complete.rar
• %FOLDER% = full path of the folder of the file, like /the/full/path/to
• %FILE% = the complete file name, like file_complete.rar
• %FILENAME% = just the file name, like file_complete
• %EXTENSION% = just the file extension, like rar

Change settings manually

If you want to change the settings manually, you have to change the file named xg.config located in your user folder:

• Windows 7: C:\Users\Username\AppData\Roaming\XG
• Linux: /home/Username/.config/XG
• Mac: /Users/Username/.config/XG

Add servers and channels

Now you have to add IRC networks and channels. The bots and packets are generated and updated automatically. If you don't know which server and channels to add, try the integrated xg.bitpir.at search or add a XDCC link.

Normally the bots will announce their pakets directly in the channel. If they are silent, you can check the option Check user versions and XG will ask the voiced users about their version. If XG detects an iroffer he will try to send xdcc list commands to get packet lists. _DO NOT_ check the option unless you know, that the bots in this channel wont announce their packets. Otherwhise you mostly will be banned!

Server / Channel Dialog

Search

You can search for packets by entering a custom search term and just hit enter. If your want to save your search, just click on the thumb button. Deleting a search works the same. The search items are working with the internal and external search and are also saved into the database. If you want to exclude words from your search you can use "-". To search for packages and exclude TS releases you could use Spiderman -TS. The size of packets can be controlled by the size box. Only packets which are bigger than the given value are displayed. If you don't want to use this feature, leave this field blank or zero.

Search

XG supportes wildcard searches to be able to search for tv shows. If you search for under the dome s02e** you will get results for all season 2 episodes from 01 to 99. Even multiple wildcards are supported: under the dome s**e** will return results for all season from 01 to 10 and episodes from 01 to 30. Because multiple wildcard searches are expensive, the results are limited to 10 seasons and 30 episodes.

The results are displayed in a table and the packets can be grouped by their bot or wildcard search. The grouping can be disabled, but you will lose some important informations. If you click on a packet icon, XG will try to download it and keeps you up to date with updated packet informations. The packet icon will match the file ending, so there are different versions.

Packet Icons

Packet Icons

Notifications

If something happens inside XG you will get a notification. This can also be shown via your browser if you allow it.

Notification Icon

XDCC Links

You can add XDCC links in the following dialog. A XDCC link must have the following structure:

xdcc:// server / server-name / channel / bot / packet-id / file-name /

The server, channel and bot is automatically added. If the server is connected and the channel joined, the packet will be requested.

XDCC Links

The server and channel are not deleted after the packet is complete, so if you don't need them anymore, you have to delete them yourself.

Extended Stats / Snapshots

XG will collect every 5 minutes some statistical data and generate nice graphs. There you can enable and disable different values to get an optimal view of your running XG copy.

Extended Statistics

This feature wont work in older browsers like the good old IE8, so do yourself a favor and use a newer one ;-)

API

XG v3 supports a REST api to control it via external tools. You can add api keys and enable / disable them.

Api

The following objects can be controlled with different methods via the api:

• servers
  • add
  • delete
  • enable / disable
  • list
• channels
  • add
  • delete
  • enable / disable
  • list
• bots
  • list
• packets
  • enable / disable
  • list
• files
  • delete
  • list
• searches
  • add
  • delete
  • list

The object name has to be entered after the /api/1.0/ path segment with the format in the wich the answer should be encoded, for example /api/1.0/servers.json. The data you want to pass to the method has to be encoded in same format. The content type must be the format, too. The Authorization header is mandatory and has to match an api key which is enabled. If the apiKey was invalid or disabled, the method will result in an 401.

The currently allowed formats:

• json (preferred)

Api methods which create or update data, always return the following properties:

• ReturnValue (int)
  • 0 - there was an error calling the method
  • 1 - everything is fine
• Message (string)
  • a helpful message if an error occurred

Delete

You can delete an object and all children.

URL

DELETE /api/1.0/[ servers | channels | files | searches ]/$guid.$format

Example

curl -H "Authorization: 615d86bb-f867-47c1-a860-ac24e09e976c" -s -XDELETE localhost:5556/api/1.0/servers/deebd412-9b16-4726-b613-7ec98e714f59.json

Return Value

{
  "ReturnValue":1,
  "Message":null
}

Get

Get a single object by its guid.

URL

GET /api/1.0/[ servers | channels | bots | packets | files | searches ]/$guid.$format

Example

curl -H "Authorization: 615d86bb-f867-47c1-a860-ac24e09e976c" -s -XGET localhost:5556/api/1.0/servers/deebd412-9b16-4726-b613-7ec98e714f59.json

Return Value

{
	"Port":6667,
	"ErrorCode":0,
	"ParentGuid":"c31aa923-b615-4d03-840d-c82357c929d4",
	"Guid":"906bfd60-b6f1-4d38-a2f4-cb9cba983a24",
	"Name":"irc.abjects.net",
	"Connected":false,
	"Enabled":false
}

Enable

If you enable servers and channels, they will be connected. If you enable a packet it will be downloaded.

URL

POST /api/1.0/[ servers | channels | packets ]/$guid/enable.$format

Example

curl -H "Authorization: 615d86bb-f867-47c1-a860-ac24e09e976c" -s -XPOST localhost:5556/api/1.0/servers/deebd412-9b16-4726-b613-7ec98e714f59/enable.json

Return Value

{
  "ReturnValue":1,
  "Message":null
}

Disable

If you disable servers and channels, they will be disconnected. If you disable a packet the download will be stopped and the file is beeing deleted.

URL

POST /api/1.0/[ servers | channels | packets ]/$guid/disable.$format

Example

curl -H "Authorization: 615d86bb-f867-47c1-a860-ac24e09e976c" -s -XPOST localhost:5556/api/1.0/servers/deebd412-9b16-4726-b613-7ec98e714f59/disable.json

Return Value

{
  "ReturnValue":1,
  "Message":null
}

Add

You can add an object. All parameters are mandatory.

If you got a XDCC link, you can use this method to add a packet and download it instantly. The server and channel are not deleted after the packet is complete, so if you dont need them anymore, you have to delete them yourself.

Url

PUT /api/1.0/[ servers | channels | packets | searches ].$format

Post Parameters for servers

• server (string): irc.rizon.net
• port (integer): 11

Post Parameters for channels

• server (string): irc.rizon.net
• channel (string): #abjects

Post Parameters for packets

• server (string): irc.rizon.net
• channel (string): #abjects
• bot (string): [XDCC]Bot
• packetId (integer): 11
• packetName (string): My.Super.Movie.mkv

Post Parameters for searches

• search (string): german -mkv

Example

curl -H "Content-Type:application/json" -H "Authorization: 615d86bb-f867-47c1-a860-ac24e09e976c" -s -XPOST localhost:5556/api/1.0/servers.json -d '
{
  "server":"irc.rizon.net",
  "port": 6667
}'

curl -H "Content-Type:application/json" -H "Authorization: 615d86bb-f867-47c1-a860-ac24e09e976c" -s -XPOST localhost:5556/api/1.0/packets.json -d '
{
  "server":"irc.rizon.net",
  "channel":"#abjects",
  "bot":"[XDCC]Bot",
  "packetId":11,
  "packetName":"My.Super.Movie.mkv"
}'

Return Value

{
  "ReturnValue":0,
  "Message":"server is empty"
}

List

You can list objects. If you want to list packets, you can controll the results.

Url

GET /api/1.0/[ servers | channels | packets | files | searches ].$format

Get Parameters for packets

• searchTerm (string) *: german -mkv
• showOfflineBots (boolean): true | false
• maxResults (integer)
• page (integer)
• sortBy (string): Id | Name | Size
• sort (string): asc | desc

The properties showOfflineBots, maxResults, page, sortBy, sort can be left blank. If you leave showOfflineBots blank, it will be filled with false and the search request will just return packets, which bots are online.

Example

curl -H "Authorization: 615d86bb-f867-47c1-a860-ac24e09e976c" -s -XGET 'localhost:5556/api/1.0/packets.json?searchTerm=mkv%20-seven&showOfflineBots=true'

Return Value

{
  "Results":
  [
    {
      ...
    }
  ],
  "ResultCount":5584
}

Results is an array containg the requestet objects.

Shutdown XG gracefully

If you want to shutdown XG, just ctrl+c the process or close the command window. You can also stop XG by using the shutdown button in the webfrontend.

Upgrading XG

If you are upgrading from version 2 to 3, you should finish your downloads and write down your servers and channels, because XG 3 is not able to load the data generated by previous versions.

If you are upgrading from XG 3.2 to 3.3 you should notice, that the db format switched from sqlite to db4o. Because of this, XG automatically transformes the db xgobjects.db into a db4o database xgobjects.db4o if it is not there already. The sqlite file can be safely deleted after the first start, but can also be keeped as backup. If you delete the db4o file, XG will start the transformation process again.

Unnecessary Files

Because XG changed some internal routines you can safely delete the following files in the config folder:

prior version 2

• XG/xgsnapshots.bin
• XG/xgsnapshots.bin.bak
• XG/statistics.xml

prior version 3

• XG/xg.bin
• XG/xg.bin.bak
• XG/xgfiles.bin
• XG/xgfiles.bin.bak
• XG/xgsearches.bin
• XG/xgsearches.bin.bak
• XG/settings.xml

Running XG

On Windows

You need at least .net 4.5.

On Linux with Mono

You need at least mono 3.x because some needed libs are running on .net 4.5 wich is not supported in earlier versions.

If you are using Debian / Ubuntu, take a look here to get newer mono packages:

http://mono-project.com/DistroPackages/Debian

Needed packets / libs

• mono-complete

Install command for Debian / Ubuntu to copy paste:

sudo apt-get install mono-complete