Website: http://www.leafseek.com Background article about how the project came to be: http://www.leafseek.com/blog/leafseek-gets-published-in-avot...
I currently work with two major non-profit genealogy groups to put their record collections online for free and easy searches, without handing them over to for-profit groups like Ancestry. So far we're at 700,000+ records online using my code, and the newest "big" addition is the Israeli State Archives agreeing to put their British Mandate period marriage and divorce certificates online via one of the non-profits using the code.
I'm planning on launching the system as a multi-tenant SaaS one of these days, once my kids are a little older. :-)
Business plan?
Data license? (Public domain, creative commons, etc)
I make everything I retrieve publicly available with a copy thrown in the Internet Archive for good measure.
No freemium, only paying clients. BUT a price discount if users allow (1) auto backup of datasets in CSV format to Internet Archive through their S3 system, even if the data is "darked" and not visible (I'm on Archive Team and I also created their "Antecedents" collection of genealogy-related web crawls: https://archive.org/details/archiveteam_antecedents) and (2) they explicitly put their data in the public domain.
Pie in the sky idea: create API for higher tier paying users to allow them to license their data directly to Ancestry, MyHeritage, FindMyPast, etc. Which would flip the pay-for-data model on its ear, allowing users to regain control of their own datasets, and for-profit data brokers to pay them for access.
To that end, I will also be rejecting VC and bootstrapping the venture. Ain't nobody going to be flipping this site or acqui-hiring that data. This is going to be done for the long-term.
Some copies of the census are digitised already, but the further you go back the less likely that is. And a lot of the information people want originates abroad (e.g. UK, Germany, Italy, France, etc).
I'm working with an organization to collect real estate parcel data for the entire US and open source it. I track every county or district that has parcel data, I built a workflow to get the data from them, manage the ingest of the data, and a separate workflow if they're difficult and I have to FOIA them for it (via muckrock.com). There is a cost, but as always, everything has a cost.
Same deal, but closer to what Distributed Proofreaders for Project Gutenberg does with scanning.
It can be done. Funds and organization would be necessary, but let's not think its some herculean task.
Scanning the images allows people to verify the transcription and to distribute the work of doing the transcription.