When we launched TweetMeme in 2008, I never dreamed of how quickly our service would take off and the incredible reach it would have.

TweetMeme was built to curate and rank Twitter links and gained 10 million monthly users in just nine months. We were the first to create the iconic (green) retweet button that was installed on 500,000+ websites, with a peak serving of 1.5 billion daily retweet buttons.

We are now in the process of shuttering TweetMeme and I wanted to share the reasons behind this decision, along with next steps to guide you.

TweetMeme was the first website to show the true power of curating news from Twitter. For millions of users, it was a homepage that showed a truly democratized view of what was popular on the Internet. Many stories broke first on TweetMeme as we cared more about the virality of a story, rather than who was saying it. When the plane landed in the Hudson River in New York, TweetMeme was the first to break the news on its homepage.

Three years on and the consumer news market has moved on and our core business DataSift has now grown to over 10,000 users, offices in 4 cities, $14m in investment and an amazing ecosystem of applications built upon it. We will be sad to see TweetMeme go, but it is no longer competitive or cost effective for us to continue to keep the infrastructure going behind it.

We have done our absolute best to ensure that the shutdown of the free TweetMeme API is as easy as possible for everyone. The shutdown of the free API which allows anyone to look up retweet counts and resolve short links to full links, is already in effect. Here is the timeline for the remaining steps as we wind down TweetMeme:

Buttons and API

Today whitelisted API will be shutdown and all whitelisted users will lose access to the /url_info end-point.

It has been two years since Twitter launched their own button and as part of the final TweetMeme transition, the TweetMeme button will over the next 24 hours be switched over to the Twitter button (the ‘web’ TweetMeme button will immediately look like Twitter buttons). Please note that this will happen automatically and no action needs to be taken from your end.

A few other important notes:

• The Retweet Button will lose some functionality when transitioning to the Twitter Button. The parameters tweetmeme_alias, tweetmeme_service, tweetmeme_service_api and tweetmeme_space are deprecated.
• In certain situations we advised to manually install the TweetMeme button using an iFrame, unfortunately this button will no longer be operational and will not be converted to a Twitter button. Sites will need to replace the embed code with Twitter’s own.
• The Image Button used within RSS feeds is being dropped.
• The Follow Button is being dropped.
• Our WordPress plugin will be switched to only use the Twitter button. As such, we strongly encourage you to upgrade the WordPress plugin when prompted from your dashboard.

Website

On the 1st October the TweetMeme website will be fully shut down.

Thanks

Thanks for your support as a TweetMeme user – it’s been a great ride. To learn more about DataSift, I encourage you to visit datasift.com. I also welcome you to stay in touch by following me on Twitter @nik and tuning into our DataSift blog.

We are happy to announce that we have added two new Windows Live Writer Plugins to our offering with the help and support of the Live Writer developer community.  Our thanks go out to Scott Lovegrove who has carefully and lovingly created the Windows Live Writer TweetMeme plugin for WordPress.com along with the Windows Live Writer Follow Button Plugin.  Whilst he was on a role and doing these new plugins he also updated the much loved Windows Live Writer TweetMeme plugin to include support for hashtags.

We hope you like the plugins and if you are a developer and want to do cool stuff with our API’s then please do take a peak at our developer documents!  Tell us about your great creations and we’ll give you the spotlight!

Today we’re releasing a major upgrade to our platform that will significantly impact parts of our API. These changes are necessary to allow our platform to scale at the tremendous rate we’re seeing and provide vastly improved reliability.

/stories/tweets

The /stories/tweets API call now includes a “uuid” (a 32 char hex string) for each Tweet returned. This UUID represents the Tweets position within the timeline for that story. The from_id parameter, which used to accept the Tweet ID of the previous Tweet to page from now accepts the UUID of the previous Tweet to page from.  If you pass the Tweet ID instead, you will get an error.

This is the most disruptive change and only affects clients that page tweets. Your underlying paging logic can remain the same, but you will need to pass our API a different value for the from_id.

/stories/recent & /stories/popular

Both the /stories/recent and /stories/popular API calls will no longer accept a domain parameter. This is due to a performance issue that we have identified and we hope to restore this functionality in the coming months.

The Stories API docs have been updated with these changes, please refer to them for more details and examples.

We are very happy today to be releasing our Twitter real-time deep link search. (quick link: http://search.tweetmeme.com ) We have been quietly building up an infrastructure that could leverage all the work we have previously done on aggregating all the links from Twitter and deep delving those links to find content and media.

We launch today with the ability to search 15+ million links which we have deep searched for content, this not only includes the text from web pages and blog posts but also videos and images.

Although the ability to keep our index totally real-time is great we also believe that the delivery of relevance is still just as important, our default search results are based upon a clever combination of newness, retweets and some other weighting factors. This means results tend to be recent and relevant but at the same time the most popular. We do also offer search by ‘Age’. Lastly you can sort the results purely by ‘retweets’ to give you the most relevant results based upon its popularity on Twitter.

Here is a quick bullet point of the key features of our search,

• Real-time search – our indexer picks up stories as they come out of Twitter
  • Boolean operators (& | – ! and grouping) – explained below
  • Sort by Best Match
  • Sort by Age
  • Sort by Retweet Count
• Filtering
  • Filter by Category
  • Filter by Media (News, Images, Videos)
  • Filter by Age (recent links first)
  • Filter by Tweets (e.g. 100+ retweets)
  • Filter by Channel (e.g. apple, start-ups)
• RSS – you can subscribe to any of the search results
• Retweet – you can retweet the links you find without leaving the site (if you are logged in via OAuth)

Some guidelines on how to use our search -> http://help.tweetmeme.com/2009/05/11/search-guidelines/

Lastly we will soon expose the engine via our API so that other developers can leverage the results.

See what the press have said already:

• TechCrunch: Tweetmeme Launches The Second Real-Time Tweet Link Search — This Hour
• VentureBeat: TweetMeme’s latest Twitter search tools take on Scoopler, OneRiot
• Mashable: Tweetmeme Shows Off the Future of Twitter Search
• Louis Gray: http://www.louisgray.com/live/2009/05/tweetmeme-launches-real-time-search.html
• CNN:  http://edition.cnn.com/2009/TECH/05/12/future.search.engine/index.html?iref=t2test_techtues

Or just search on TweetMeme

So today is my first day as the community manager at TweetMeme and as such I thought it would be a good opportunity to introduce myself and what I’ll be doing here over the next few months.  So here goes…

First things first I’ll be looking over the comments from our current users and identifying any key areas where the site could be drastically improved so get your suggestions into us via the forum as I’ll be looking there to see what areas are your pain points and what features we are missing that should really be there.  I’m also into making things as simple and usable as possible so if I see usability issues I’ll be picking up on them to make your job easier!

Secondly I’ll be attending The Next Web conference and will have my laptop with me so that you can see TweetMeme in action and at it’s best.   As such do come and find me and I’ll show you what we can do with the site!  Honestly there’s more to it than just the front page… I wonder how many of you are using the tweetmeme wordpress plugin on your sites or the buttons.  Also if you’re a developer and are interested in working with our API then do take a look at it, we are looking for feedback on this as it’s something we are hoping for developers to get really into!

And finally some random facts about me…

Background: Formerly I worked at Cardinal Health as a software engineer doing a  lot of .net programming and software development in and around this area.  I’m also the founder of the Girl Geek Dinners and that is where my community experience comes from.  I’m known for being a little unconventional in my approach to things and have a passion for mobile devices… so thanks to Nick (aka the Boss) I am now fully connected with an IPhone!  So that means you can find me on Twitter a little more often than you used to!  If you want to find me on twitter my account is @girlygeekdom and I have a couple of blogs… my group blog is GirlyGeekdom and my personal one which I’ve not put much on recently is Flirting with the Future… I thought it was a fun name for a blog!  You’ll also find me blogging here and over on the Fav.or.it blog too! And yes I will be cross posting every now and again, but hopefully not too often!

The twitter firehose is a LOT of data, TweetMeme spends it’s life sifting through finding all the URL’s resolving millions of shortened URL’s (e.g. http://bit.ly, http://tinyurl.com) in the process. Once resolved we go grab the title of the story, and also do a lot of clever stuff to get images/videos + the body of text (i.e. the description of the story).

We thought it would be useful to expose this data via a set of API calls which developers could use to improve their twitter applications who do not wish the massive burden of processing all the data that TweetMeme does.

From Today we launch 3 simple methods that expose a small part of our dataset, these will hopefully be followed by more that allows data-mining around the unique resolved-stories dataset we have.

If you want to dive straight into the technical description please go visit our API page

The three calls we have exposed so far are,

• url_info – This allows developers to take any URL posted to twitter and get back the title, category, tweet count, date posted and the resolved URL.
• recent – Grabs the most recent stories from Tweetmeme
• popular – Grabs the most popular stories from Tweetmeme

If you wish to get updates on further developers follow tweetmemedev