Being able to write files to a filepicker url - which ends up in a users cloud account - is very powerful stuff. We're using it to automatically sync zip files that take a long time for us to process. Having a very easy to use layer of abstraction for all these cloud APIs is brilliant.

This has probably been discussed before, but why the huge gap in price between free and "Pro"? Why not make "Free" a max of 500 uploads/month and a "Starter" tier that costs $10.00/month for 5,000 uploads?

My guess is it's more valuable to have the people with 500 < x < 5,000 uploads/month comfortably try it out for free, and possibly upgrade, than debate whether or not it makes sense to spend $10.

Accurate. Also, the free plan is very generous at the moment as we want to reward our earliest users. A small tier might make a lot of sense later on.

whoa. you can read and write on their urls? pretty slick as it actually looks like a js file system.

It's super interesting; javascript is okay, but any shortcomings seem to be solved by other people.

Inconsistencies and ease of use- Jquery Filesystem- Filepicker Code organization - Backbone etc.

I really wanted to use your service in the past but I couldn't justify paying so much just for image conversion.. I see it now comes as standard! Thank you!

From the documentation it appears that you can write to any file if you have filepicker.picked it. Is this true even if the file was outside the browser file system sandbox? For example I select /home/me/myfile.txt ? Assume the file has write permissions for anyone. Can you write directly to myfile.txt? Or is a push to S3 + download involved.

There are some files where it is not possible to write back. Local files is a good example of this. We throw a 409 HTTP error for files like that. You can of course write to the s3 copy.

Saw that you guys changed your pricing to per file (specifically, it says per FPURL created). Does that mean if I have a user upload a file, and then convert that file by cropping/resizing it, each time I convert it that also counts as a "file" toward my monthly quota? Or are conversions of uploaded files "free"?

we don't charge on conversions of uploaded files. Let me know if you have more questions at adass at filepicker.io and I'd be happy to help.

EDIT Ah. I misunderstood. While we don't charge on conversions, we do create a new filelink for each conversion which is counted under your quota.

If you're on the free plan and think this will hurt you ability to test us out, let me know. If you're on the pro plan and it's hindering your ability to use us, also let us know. We'll work something out.

I'm reading through the docs, but not quite understanding the architecture. Is Amazon S3 specially privileged over the other cloud services, or is it just being used as an example? Do all file requests mediated by the Filepicker.io servers, or can one go straight from browser to 3rd party storage?

Hi there.

Currently S3 is the only support backend storage. More are on the way. We take requests from customers seriously and so far we've heard rackspace, google, and dropbox. Feel free to send me an email at liyan@filepicker.io with your preference.

All files are mediated by fp.io. This lets us do interesting thing like implementing tcp/ip on tcp/ip. That is, we chunk up large files into manageable bits, retry, and reassemble on our servers.

Any guidance on upgrading to the V1 API? I'd love to use the file extension limits instead of the MIME types (which can't target some files like *.ipa)...

Should be pretty straightforward. If it's not, shoot me an email and tell me where I messed up ;D

getFile is now replaced by the more explicit calls: pick and store.

If you're using the widget, data-fp-extension is now available.

Nice work FP.io team!

Thanks. The support of the community is really encouraging and keeps us going. Always be shipping. :D

More great stuff you guys!

Where's creat?

Being able to create arbitrary files without asking the user gets into some nasty security issues, so we went with "export" instead. Once the user selects where they want to save the file to, you can then perform .read's and .write's on it

would this work in node.js?

No reason why not. Give it a spin and let us know how it goes. Also happy to help if you run into issues.