From arXiv's website:
Started in August 1991, arXiv.org (formerly xxx.lanl.gov) is a highly-automated electronic archive and distribution server for research articles. Covered areas include physics, mathematics, computer science, nonlinear sciences, quantitative biology, quantitative finance, statistics, electrical engineering and systems science, and economics. arXiv is maintained and operated by the Cornell University Library with guidance from the arXiv Scientific Advisory Board and the arXiv Member Advisory Board, and with the help of numerous subject moderators.
arXiv has a form of moderation. Users must be "endorsed" by people who have themselves posted papers on the arXiv. (more info) Moreover, a team of moderators is able to reclassify or remove content from the arXiv, for a number of reasons: unrefereeable content, inappropriate format, inappropriate topic... (more info) It also implements some automatic checks for content overlap between papers.
In some fields (e.g. math) it is the standard way of releasing a new paper into the world. At the moment there are 1,369,674 papers on arXiv.
arXiv has a plan to ensure its sustainability and make sure that papers uploaded to arXiv will not disappear from the internet should anything bad happens. There also exists several mirror websites should it even go down.
From viXra's website (in the footer):
ViXra.org is an e-print archive set up as an alternative to the popular arXiv.org service owned by Cornell University. It has been founded by scientists who find they are unable to submit their articles to arXiv.org because of Cornell University's policy of endorsements and moderation designed to filter out e-prints that they consider inappropriate.
ViXra is an open repository for new scientific articles. It does not endorse e-prints accepted on its website, neither does it review them against criteria such as correctness or author's credentials.
It was founded by a disgruntled physicist named Philip Gibbs. It is unclear who funds the website today, and it is not even clear who actually runs it today if you just read the website. Compared to arXiv, there is little moderation (according to the FAQ, a paper can e.g. be removed if it's not science/math, if it's plagiarism, if it's obscene and so on, but not if it is obvious nonsense unless it is computer-generated nonsense) and people can submit papers anonymously or under fake names.
At the moment, there are 22,596 papers on viXra.
viXra has no sustainability plans. If whoever runs it decides to take it down tomorrow or their servers catch fire, I'm not sure what would happen.
Warranted or not, it has a reputation of being an alternative to arXiv for cranks and to host a lot of junk science, fake proofs or even outright nonsense. It is not recommended to submit something to viXra if you wish to be taken seriously, because the reputation of viXra would probably taint your paper's reputation by association. (Which is not the same thing as saying that arXiv "bestows credibility" on papers found there. It's rather viXra which bestows a bad reputation.)
arXiv.org is a multi-institutional e-print repository funded by universities and other research organisations costing hundreds of thousands of dollars to run. It's purpose is rapid dissemination of papers and to provide a permanent open access archive for the professional research community. Submission of papers is in principle open to anyone but those outside the academic world are hampered by a need to be endorsed by an insider. Even when endorsed, papers that do not come from a respected institution are subjected to a moderation process that can result in them being delayed, rejected or reclassified in a generic subject category that may limit how well they are disseminated. The moderation process can sometimes affect professional researchers too if their paper is flagged up by a keyword filter. (Note that it is not known if the institution is used directly by the filter as a criterion but the effect seems to be as stated)
viXra.org is an independent repository intended to cater for those who find it difficult or impossible to submit to arXiv.org or who do not like having their work delayed. It does not require endorsement and there is no moderation process. Some papers may be removed for technical reasons but there is no minimum quality standard. Like arXiv.org its purpose is rapid dissemination and open archiving.
viXra.org is funded by a small group of administrators who run it (including myself) The costs are now so low that funding is barely worth mentioning. It main site runs from a standard GoDaddy web hosting service with another similar mirror site elsewhere. The whole process is almost entirely automated with intervention only being required to respond to occasional emails or technical hiccups. Although its future is not as well assured as arXiv.org it does have a decentralised architecture and replicated shared cloud storage that should make it possible to pass control from one administrator to another when necessary. In my opinion there is a much bigger problem of loss of research papers as small journals come and go but nobody worries about that. viXra already survived an incident where its main server died with a disk failure by quickly re-pointing the DNS to its mirror site. At the time it was using a dedicated server but it was subsequently moved to a virtual server that provides far better robustness at lower cost.
The typical quality of papers on viXra.org is much lower than that of arXiv.org because of its open submission policy and because many of its users are independent researchers. However, viXra does have a core of good respectable papers from a diverse range of authors. Many people who look at viXra.org underestimate the amount of quality research it contains because the technical and diverse nature of papers mean that the good ones are only understood and appreciated by specialists while anyone can recognise the low quality content.
Some people advise not to use viXra.org because authors who submit there are "not taken seriously." This misses the point that the purpose of a repository is dissemination and archiving. No repository can bestow credibility on the papers it contains. This can only be done through peer-review, citations, experimental verification etc. Even arXiv.org does not claim to provide peer-review and most of its papers are only checked for quality by a dumb automated filter. This is why papers on arXiv.org are not automatically considered reliable sources in Wikipedia for example. Even though there is a significant difference between average quality levels on arXiv and viXra, it is still more reliable and easier to judge the quality of a paper in a repository by looking at it rather than by looking at where it is stored.
Some academics strongly disapprove of viXra.org. I think the main reason for this is that they tire of receiving unsolicited emails from "cranks" which they also associate with viXra. The situation is ironic because in fact it is the arXiv endorsement policy that is implicitly encouraging people to try to contact academics in order to get recognition and an endorsement. viXra.org discourages this because it is better and safer to make research available publicly than to send it to people privately. However, viXra.org has no means to police the habits of the minority who ignore this advice.
viXra.org does have some popular features that arXiv.org does not including commenting, download statistics and easy withdrawal or removal of old versions.
arXiv is a respected repository for physics and math preprints. In some fields of physics, it is actually the primary venue through which new papers are read.
viXra is a site for people, almost exclusively cranks, who cannot or will not put their material on arXiv. Don't use it.
viXra.org (arXiv spelled backwards) was founded in response to the many reports of arXiv.org's censorship, victims of which include Nobel Prize laureates and first-class university professors. See their "Why viXra?" page.
In reporting these case histories we do not imply that we are embracing, supporting or endorsing, in any way, the particular ideas, work, principles, or religious and personal beliefs of these scientists. Our aim is to include a wide range of victims, scientists from all over the world, who range from Nobel Prize winners to relatively unknown scientists, those with numerous publications in first rate journals, to those with a few publications in second rate journals, from professors in first class universities to humble scientists working outside academia, from rigorous scientists to those exploring the fringes of science.
The New Endorsement System: Merely a Smoke Screen
In January 2004, arXiv.org introduced an automated endorsement system which was touted by operators of the system as an answer to complaints about the archive's restrictive policies. However, there have been cases where blacklisted scientists have attempted to use this supposedly objective endorsement system and the outcome has been overriden to ensure that their works are still blocked.
Another drawback is that most endorsers are fearful of endorsing novel papers because they worry that they could lose their endorser status if their decision runs counter to the tastes of the arXiv moderators. In fact, the arXiv.org website states that the archive administrators reserve the right to revoke any endorser's endorsement status. When one blacklisted scientist's paper was rejected, the reason given by the archive moderator was: "No legitimate endorsements from current users." When the scientist then asked "will you guarantee that persons who endorse me will not be victimised by having their arXiv privileges withdrawn?" no substantive response was forthcoming.
The scientist concluded from this that they had refused to guarantee that anyone who endorsed him would not be punished for doing so. Thus one could surmise that the endorsement system is merely a facade to give the false impression that the electronic preprint archive operates in a fair manner.
Another important difference that isn't outlined in the other answers above:
arXiv supports HTTPS, so you can't be spied on what you read or upload or have content injected into your connection (a JS miner for example).
viXra doesn't seem to support HTTPS, which means that all of your identifying information that you use to log in can be tampered or stolen by anyone who is able to watch your traffic or viXra's server traffic.
Note that non-HTTPS for a website which requires login forms is unacceptable in 2018 especially since there are now free SSL certificates such as Let's Encrypt.
Moreover, according to the Mozilla Observatory:
- arXiv has a "D" rating https://observatory.mozilla.org/analyze/arxiv.org
- viXra has an "F" (!) rating https://observatory.mozilla.org/analyze/vixra.org
tl;dr: arXiv cares more about your security than viXra.