This post is a brief update about the 
STATUS: Currently upvoting 374 unique users with 24 voting accounts that collectively control over 10000 SP. Closed invite-only trial mode.
@we-are?The 
It will allow communities to access voting power without having to set up any technical infrastructure. There will be simple steem memo based system for managing memberships. Community levels will be expected to ensure the posting quality of their members and to remediate, rehabilitate or eject members that are posting poor quality work.
Donors can target their SP-delegations at a level of granularity that suits their policies. There will eventually be reporting on the effectiveness of their donations.
The core feature, automated upvoting of posts by members from communities, is working. The voting accounts arrange themselves in a flexible manner so that many communities are served by more than one voting account. For example, Nigeria communities are upvoted by 
The voting account 
The 
then I will consider adding your community to the trial. Contact me on discord.
I originally paid a programmer to create the first version of this bot. His model worked very well and I consider the money well spent, however it relied on a Steem Database service. I spent the first few weeks changing the system to work with the steem blockchain instead. This version only supported a single voting account and all config was in a JSON file.
My first version of the bot would check every voting account every seventeen or so minutes, and then, for each voting account go and fetch the most recent posts for each
The bot was originally coded with a "philosophy" of javascript that, while still valid, was a bit outdated. My own additions used a different philosophy until I decide we should migrate the code, gradually, towards using a Promise() based philosophy. That meant the code base was horrible. The config was also placed into an SQLite database too.
It became clear that I needed to rebuild the bot from the group up to use a more scalable architecture. I changed to using dsteem instead of steemjs and found it much easier to follow. The SQLite config database is retained. The most significant architectural change is to streaming the whole blockchain, filtering the stream for events (e.g. posts) and then passing these off to be voted upon. This is much more scalable than continually polling for data about members that might have gone inactive. The bot is quite resilient because it persists crucial data to LevelDB.
Right now, the new bot is not quite the full functionality of the old bot. However, with the new architecture, it is much easier to add on new features.
The main thing is that I expect to be able to gradually expand the use of the bot (see terms above) to test its scalability. The next significant set of features I will work on are:
I don't have timelines for these changes. I will fit them around my work schedule.
@eturnerx