There's nothing as nice as receiving unsolicited customer testimonials and there's nothing as nice as being able to get them on your Web site quickly. In the old days testimonials came in via e-mail, if you were lucky. These days, if you're doing your job right you are likely to be receiving plaudits unbeknownst to you on a variety of social networks. A variety of tools such as Whos Talking, Samepoint and Social Mention exist which let you easily search across multiple social networks and blogs for mentions. The latter even gives you a neat little Flash widget that you can embed on your Web site so that visitors can see those mentions fly past.

But what if that stream contains some ... ahem ... less than complimentary mentions? That's where Habit Stream comes in, it's aim - to let you search multiple sources, choose only the nice things that people are saying about you and get them displayed on your site with the minimum of fuss.

The system is simple enough to operate:

• Set up a "stream" that will containing a series of search terms.
• Put the the terms into your stream.
• Pick the items you like from the search results.
• Create a "Broadcast" to display the results embedded on your site.

The "Broadcasts" come a three varieties so far, a customisable Flash widget, a "mini-site" option that can be embedded into the main site or linked to and a full API for developers with more ambitious plans for integrating the comments into their sites.

Setting up searches and selecting items is simple, but the results need better filtering options - coming soon, apparently.

Some aspects of the service are clearly still a work-in-progress and not fully baked. 'Out of the box' it only searches a relatively small number of sources, although additional RSS and Atom feeds can be added to a search. Comments on YCombinator's Hacker News suggest that Habit is working hard to expand the selection. But the biggest problem is the sheer amount of noise that the searches currently include. The results tend to be cluttered with rubbish that cannot be adequately be filtered out since there isn't support for the full range of boolean search modifiers yet. Thus It is impossible to restrict results to only those posts that contain the words Chris Noble and the words "truly excellent" (something I've been Tweeting regularly - simply for the purposes of search testing, you understand).

Habit founder Devin Hunt says says the firm is working on a "robust filtering system" which it is hoped with solve some of these problems. But, to my mind the thing needed to make the service really useful is the provision of sentiment analysis, which which, even if not perfect could help bubble positive comments towards the top of the results. Integrating something like OpenAmplify, Calais or the like into the offering would make it more compelling. As an aside you can have fun seeing OpenAmplify sentiment analysis in action at Tweet Sentiments.

At the moment then, the service has potential, and the Broadcast module combined with the pick-and-choose Curating function certain a convenient way to get the right live comment onto a site. It's particularly well suited to PR agencies who can use it to seamlessly curate the content on their clients' sites... and can pass the cost along. Smaller, more tech-savvy operations are likely to be content to spend a few extra minutes needed to use the free search tools out there, copying and pasting the results into their sites content management system. A 30 day free trial is available, the ability to use fully customisable white-labelled embedding starts from $99 a month.