Greplin’s (YC W10) Social Search Opens Its Doors To All | Hacker News Reader

Greplin’s (YC W10) Social Search Opens Its Doors To All

techcrunch.com

66 pointssmanek21 comments

This sounds great, but I'm a little wary of giving a single service (especially a small new startup) access to so much private data.

They do get points for requiring HTTPS and using OAuth instead of asking for passwords, though.

I think I'll start with the low risk accounts like Twitter and LinkedIn for now.

I like what Greplin is trying to do, but I wont sign up for an account. I am uncomfortable authorizing one service to access so many of my accounts.

Edit: Greplin doesn't request usernames or passwords. I updated my comment accordingly.

We don't actually store usernames and passwords. We're all OAuth, all the time, and you're welcome to revoke access whenever you like. We'll also delete all of your data as soon as possible when you ask us to.

Do you encrypt data? If your servers are compromised, will someone have access to all of my emails?

Even if that is the case, I might be more comfortable using Greplin if I knew more about your security practices. What procedures are in place to prevent my emails from being stolen?

Edit: I hate being the one to criticize an incredible idea that is the result of hard work. I feel obligated to add that I want to use Greplin and I probably will use it for Facebook and Twitter. (Facebook's built-in search is useless.) I really hope there is a way that I can use this for email too.

Personally, I find GMail's inbuilt search hard to beat. Its almost perfect. But I've used Greplin for Facebook (that's the only thing that I'm using Greplin for), and I find it to be an amazing service.

Doesn't Facebook already have a search feature?

Its fairly limited, feature-wise.

Do you store full copies of all data? What I mean is, if someone breaks into Greplin, can they effectively read all of my email assuming I've synced with Gmail? Or do you just index the data and reference sources using URLs?

Even if they just have an index without the full copy it's not that hard to reconstruct a version similar to the original just from the index, as in an inverted index you typically do not only store the documents of a term, but also the word number within the document. However, it's not possible to restore the original version exactly, due to things like stemmers.

often in text indexing the original document is kept for things like snippet generation.

Every service I connected (Twitter, LinkedIn, GMail, Dropbox, Facebook) uses OAuth, so Greplin never sees your passwords.

A locally-hosted Greplin-like service is an idea I've been kicking around for a while and slowly prototyping -- my angle is more on the archival side, but once you have the data search is a natural step. Another interesting attempt down these lines is the (just recently announced) Locker project https://github.com/quartzjer/Locker (being written by the guy who created XMPP).

A locally-hosted Greplin-like service is already available. Checkout http://www.cloudmagic.com

I signed up, and Greplin finally finished indexing my data (it took a few days). However, a big minus for me is that Greplin doesn't search Gmail chat logs, and it doesn't disclose that until I searched through the FAQs. I hope they're able to add that soon.

My indexing only took a couple of hours. I also quickly discovered to my dismay that GMail chat logs aren't indexed.

Great concept and execution! I've had the same idea for years but thought Google would launch this sooner or later. Love that it was launched with a business model as well.

Some feedback:

* I would be able to search local files as well. Dropbox is a great start but I want all of it ;)

* The UI have some neat features (love the interactive graphics on the upgrade page!) but overall it has to be improved a lot. Better graphics, better search result page, typography and especially all the buttons.

* Another way of monetizing the idea is to have partnerships with backup services. Affiliate deal or whatever. Since the user want the content searchable, they probably want backup of it as well.

Nice integrations.... Any plans to include web search? I don't see myself adopting the behavior of having to think to go to google for web search or greplin for non-web search. Search is search. If I have to go somewhere else, I imagine I'd likely visit the actual destination (FB, Twitter)... Thoughts?

I can envisage a chrome extension that would embed greplin results within your organic google results...

Infact i have already asked for API access, heh.

Just signed up for every service except Facebook. While Facebook would be probably the most useful service to have enabled, it's really hard to justify turning on the Facebook applications platform for any reason.

Good luck for them, the service seems nice.

The "autocomplete your life" slogan at the end of the video puts a smile on my face

Reply on news.ycombinator.com