Looks good! Please add either fastmail api or generic imap crawling to your list. :-) Perhaps you could also add support for some kind of generic webhook, or provide an sdk, so that we can add our own custom datasources quickly and easily? You could even licence the sdk in such a way that any implementation was required to be republished...
Thank you for the feedback! IMAP is high on our list of TODO integrations. Custom datasources is also something we want to eventually build support for, as we get more widely adopted.
We've started off focusing on search but will evolve over time to tackle other productivity drains and busywork in the modern office (think excessive meetings, form filling, timesheets, updates - basically everything we hated when we worked 9 to 5!). We plan to keep the search experience free and (hopefully) add enough value with these add-ons that people and companies consider paying us for them
Thanks for the tip on JMAP. I'm not familiar with the protocol but will look into it
2000s, but yes, I miss having that. Though Raycast is a million years ahead of this product, and allows you to do a search that searches gmail on web etc, or if you email client supports apple script that can do there as well, and of course lots of chatgpt integrations and gemini for searching outlook, gmail etc.
I found Spotlight's integration with Drive hit and miss, and Gmail didn't work at all for me because I don't use the macOS Mail client. I also have multiple accounts (work, personal) and couldn't find a way to search across all of them. There are also other services which folks on this list and others have asked for (Teams keeps coming up for example) which lack any integration.
This is very, very unimportant - but the build in your demo video on the site has a typo in the application name, "Selectirc" instead of "Selectric" (seen in the menu bar)
Reminds me of a modern macOS version of “Everything” for windows. Love the concept!
It would be interesting if there was an option to add additional devices, such as a MacBook or iPhone, and search across all of them at once.
Alternatively if you want to preserve isolation perhaps a mobile app would suffice.
If you built an interface that we could plug other platforms into, I could see that being helpful, akin to what raycast has done with their extensions.
1) Additional devices (not to mention Windows support) is high on our todo list. We have some ideas on how do it in a privacy safe way. Our plan is to pick that up once we know the baseline search experience works well for people.
2) Interface to onboard other platforms - great idea! Out of curiosity, what other platforms/services would you personally want to see added?
Re: proof point, what you're trying to show is that this won't steal passwords or data or be sold to someone who will, so you have to put some reputations behind it. The best thing would be to put your team on the site -- right now there are zero people on the site with their names behind this. Put who they are, where they worked before, and have them refer to this project from their LinkedIn/Github/etc. so it's a two-way link -- then there's reputations at stake. Then if you can also get some people with larger reputations than yours to believe in you and put their reputations behind it the same way, via an advisory board etc, that's even better.
Your privacy policy and terms are not bad, in that they address your specific software and describe an opt out for collecting aggregated data. I'd proofread, since typos like "Users also have the option of opting out of this collect when" make it sound like you haven't had lawyers look at your privacy approach. And "We may collect information about how you use the Service, such as search queries" sounds ominous, maybe qualify that up front.
Finally, this is when people start asking what your plan is for monetizing, since "sell the data" is the default plan. Having a good answer to that on the site would be great.
Great feedback Jack, thank you for taking the time to review and write this! We'll address these points on the site as well. For folks following this thread, this is me: https://www.linkedin.com/in/adityawasudeo/
On monetization, it's still early days but one thing we're clear about is not building an advertising product or adopting any business models which rely on accessing or trading user data. Our philosophy is more on the lines that we'd rather not have any access to user data or identity to avoid legal/regulatory complexity!
Our ambition is to incrementally tackle the various time sinks knowledge workers face at work (finding information, paperwork, meeting hell etc.) - basically the same nonsense we personally endured for years in Big Tech! If we do that competently I'm optimistic that companies and people will value the productivity gains enough for us to make a half decent living.
You have no information about where you are, who you are, etc. I might have installed this 15 years ago, but I would never use an app where it has access to my emails when I know zero about the company.
You start off saying "I" and then continually start saying "we". This might be some weekend project where you haven't gotten to any of that, but to trust something to read my emails, which contain so much personal information, is just a no-go.
I wonder how these apps secure the “indexed” data on device. These on-device databases are probably ripe for exfiltration by malware to get all of your juicy details in a neat, formatted table/database format.
Got it, Thank you. It's quite the hallucination, we'll continue looking into ways to keep the model on task! Please keep the feedback coming, these small models behave a bit differently from their large, cloud hosted counterparts.
I took a look at Alfred Workflows when we were first starting to build this a few months back but couldn't figure out a way to implement the local crawling and indexing functionality esp. for online services. The alternative would have been indexing and storing data on the cloud but we wanted to prioritize privacy. We wouldn't use a service which crawled our personal data and stored it in the cloud so it didn't seem fair to ask anyone else to do so!
I'd be okay with installing both and just having a callout to the program via alfred :). It'd be a useful way to trial. Totally understand if not worth it though, looking forward to seeing this grow!
Thank you! We're different in a couple of ways:
1) In addition to local files, we also support your cloud accounts - Gmail, Drive, Outlook, OneDrive, Slack, Dropbox (for now). IIRC Raycast file explorer currently crawls your local computer but please correct me if I'm misinformed
2) I touched upon this in another reply. All data we crawl and index is stored on your local computer alone. We also run our AI models locally so no data ever needs to leave.
Nothing needed on the Slack side. Login with your Slack credentials through the OAUTH flow, grant access to your workspace and it'll start crawling. Drop me a mail on adityaATwasudeo.com or discord (https://discord.com/invite/RbYYU6mU) in case you run into any problems.
Very weird. It was working a couple of hours back when I tested it. Second time this is happening to me with Discord. Here's a fresh link and I'll update the site (after troubleshooting this a bit more) - https://discord.gg/R63PTQ6CFh
Do you mean iCloud mail? We looked into it but Apple unfortunately does not support an API for it. Another potential option could be using IMAP but we haven't looked at it just yet.
No. I mean the Mail app itself. It can be used like a lite version of Outlook.
The app can be used as a client for standard IMAP/POP mail servers. It also has GMail integrations.
The mail is usually downloaded into local mbox files.
Ah understood. I'm not familiar with the mbox format but from a quick search it looks like they're text files so should be easy enough to build support for. Let me test this and get back
I thought the same, then checked the marks, and found them expired.
If the original owners are so sure the names are dead they let the marks expire, well, "SELECT" electric is pretty good, and fewer people than ever have any exposure to the original, so why not?
Looks good! Please add either fastmail api or generic imap crawling to your list. :-) Perhaps you could also add support for some kind of generic webhook, or provide an sdk, so that we can add our own custom datasources quickly and easily? You could even licence the sdk in such a way that any implementation was required to be republished...
Thank you for the feedback! IMAP is high on our list of TODO integrations. Custom datasources is also something we want to eventually build support for, as we get more widely adopted.
I think custom datasources would massively help drive adoption - we'd deploy across the org AND across our client base if this was an option.
Interesting, what sorts of custom datasources would you connect if you could do so?
What is your monetization plan if it’s currently free? Looks great, but bills need to be paid!
Please consider eventually supporting Fastmail’s JMAP protocol (which they designed as a successor to IMAP).
Edit: https://jmap.io/
We've started off focusing on search but will evolve over time to tackle other productivity drains and busywork in the modern office (think excessive meetings, form filling, timesheets, updates - basically everything we hated when we worked 9 to 5!). We plan to keep the search experience free and (hopefully) add enough value with these add-ons that people and companies consider paying us for them
Thanks for the tip on JMAP. I'm not familiar with the protocol but will look into it
In the 90's there was 'Google Desktop' which can do such things.
Still miss it.
2000s, but yes, I miss having that. Though Raycast is a million years ahead of this product, and allows you to do a search that searches gmail on web etc, or if you email client supports apple script that can do there as well, and of course lots of chatgpt integrations and gemini for searching outlook, gmail etc.
Same here!
Reminds me of Greplin / Cue, which Apple bought
https://techcrunch.com/2013/10/03/cue-acquired-for-over-40m-...
true. It's too bad that greplin's vision hasn't made it into MacOS yet, though spotlight is better than it once was.
I'm surprised spotlight, with the exception of Slack, doesn't just do this fine enough. What was so frustrating?
I found Spotlight's integration with Drive hit and miss, and Gmail didn't work at all for me because I don't use the macOS Mail client. I also have multiple accounts (work, personal) and couldn't find a way to search across all of them. There are also other services which folks on this list and others have asked for (Teams keeps coming up for example) which lack any integration.
I _really_ like this idea.
Is the source code available somewhere?
This is very, very unimportant - but the build in your demo video on the site has a typo in the application name, "Selectirc" instead of "Selectric" (seen in the menu bar)
Great catch, Thank you! We'll fix that.
Reminds me of a modern macOS version of “Everything” for windows. Love the concept!
It would be interesting if there was an option to add additional devices, such as a MacBook or iPhone, and search across all of them at once.
Alternatively if you want to preserve isolation perhaps a mobile app would suffice.
If you built an interface that we could plug other platforms into, I could see that being helpful, akin to what raycast has done with their extensions.
@vinnymac, Thanks for the feedback
1) Additional devices (not to mention Windows support) is high on our todo list. We have some ideas on how do it in a privacy safe way. Our plan is to pick that up once we know the baseline search experience works well for people. 2) Interface to onboard other platforms - great idea! Out of curiosity, what other platforms/services would you personally want to see added?
Can you provide a demo output of AI capabilities by using Enron emails dataset and how AI answers those questions asked by lawyers in that case?
Can anyone vouch for the privacy and overall trustworthiness of the team? I really love this concept and want to install it immediately!
I can vouch for the fact that we don't have access to any of your data! From your perspective, what's a good proof point you would like to see?
Re: proof point, what you're trying to show is that this won't steal passwords or data or be sold to someone who will, so you have to put some reputations behind it. The best thing would be to put your team on the site -- right now there are zero people on the site with their names behind this. Put who they are, where they worked before, and have them refer to this project from their LinkedIn/Github/etc. so it's a two-way link -- then there's reputations at stake. Then if you can also get some people with larger reputations than yours to believe in you and put their reputations behind it the same way, via an advisory board etc, that's even better.
Your privacy policy and terms are not bad, in that they address your specific software and describe an opt out for collecting aggregated data. I'd proofread, since typos like "Users also have the option of opting out of this collect when" make it sound like you haven't had lawyers look at your privacy approach. And "We may collect information about how you use the Service, such as search queries" sounds ominous, maybe qualify that up front.
Finally, this is when people start asking what your plan is for monetizing, since "sell the data" is the default plan. Having a good answer to that on the site would be great.
Great feedback Jack, thank you for taking the time to review and write this! We'll address these points on the site as well. For folks following this thread, this is me: https://www.linkedin.com/in/adityawasudeo/
On monetization, it's still early days but one thing we're clear about is not building an advertising product or adopting any business models which rely on accessing or trading user data. Our philosophy is more on the lines that we'd rather not have any access to user data or identity to avoid legal/regulatory complexity!
Our ambition is to incrementally tackle the various time sinks knowledge workers face at work (finding information, paperwork, meeting hell etc.) - basically the same nonsense we personally endured for years in Big Tech! If we do that competently I'm optimistic that companies and people will value the productivity gains enough for us to make a half decent living.
You have no information about where you are, who you are, etc. I might have installed this 15 years ago, but I would never use an app where it has access to my emails when I know zero about the company.
You start off saying "I" and then continually start saying "we". This might be some weekend project where you haven't gotten to any of that, but to trust something to read my emails, which contain so much personal information, is just a no-go.
Good luck on the launch! Do you plan more connectors? Would love to see Teams, onedrive etc
Absolutely, we already support Onedrive and will look into adding Teams support next. Any other connectors which you'd find useful?
I wonder how these apps secure the “indexed” data on device. These on-device databases are probably ripe for exfiltration by malware to get all of your juicy details in a neat, formatted table/database format.
Thanks for bringing this up. We store the index encrypted on disk with a randomly generated key. The key is itself stored in the Keychain.
I searched for the name of the company I'm talking with in my next meeting and it gave me a recipe for rice? Company name was unrelated to rice!
If you have a minute, are you able to share a screenshot at adityaATwasudeo.com? I'd love to take a look at this.
Edit: We also have a discord if that's easier - https://discord.com/invite/RbYYU6mU
Sent!
Got it, Thank you. It's quite the hallucination, we'll continue looking into ways to keep the model on task! Please keep the feedback coming, these small models behave a bit differently from their large, cloud hosted counterparts.
Could this be a plugin for Alfred? https://www.alfredapp.com/
I took a look at Alfred Workflows when we were first starting to build this a few months back but couldn't figure out a way to implement the local crawling and indexing functionality esp. for online services. The alternative would have been indexing and storing data on the cloud but we wanted to prioritize privacy. We wouldn't use a service which crawled our personal data and stored it in the cloud so it didn't seem fair to ask anyone else to do so!
IMO this is your business model. Build an Alfred clone that is AI first. It has a potential to be a nice lifestyle business.
Mac users still buy software and I’ve wanted to buy something like this for years.
I'd be okay with installing both and just having a callout to the program via alfred :). It'd be a useful way to trial. Totally understand if not worth it though, looking forward to seeing this grow!
This - and Raycast [0]
[0] https://www.raycast.com/
There is a nice plugin for Alfred in Ruby that searches Google Drive
If I'm not mistaken, selectric is the name of an Emacs package that adds sound(s?) when you type from some IBM (?) Selectric typewriter.
It is! It mimics the sound of the Selectric typewriter (itself a classic product design masterpiece IMO)
Congrats on the launch! this looks interesting, but how is it different from raycast's search / alfred?
Thank you! We're different in a couple of ways: 1) In addition to local files, we also support your cloud accounts - Gmail, Drive, Outlook, OneDrive, Slack, Dropbox (for now). IIRC Raycast file explorer currently crawls your local computer but please correct me if I'm misinformed 2) I touched upon this in another reply. All data we crawl and index is stored on your local computer alone. We also run our AI models locally so no data ever needs to leave.
How does the slack integration work? Do we need anything on the Slack side?
Nothing needed on the Slack side. Login with your Slack credentials through the OAUTH flow, grant access to your workspace and it'll start crawling. Drop me a mail on adityaATwasudeo.com or discord (https://discord.com/invite/RbYYU6mU) in case you run into any problems.
FYI your Discord invite link in the bottom right corner is invalid!
Very weird. It was working a couple of hours back when I tested it. Second time this is happening to me with Discord. Here's a fresh link and I'll update the site (after troubleshooting this a bit more) - https://discord.gg/R63PTQ6CFh
Turns out you have to jump a couple of hoops to create a Discord link which never expires. Should be fixed now, thanks for flagging!
Any plans to add support for the macOS Mail app?
Do you mean iCloud mail? We looked into it but Apple unfortunately does not support an API for it. Another potential option could be using IMAP but we haven't looked at it just yet.
No. I mean the Mail app itself. It can be used like a lite version of Outlook. The app can be used as a client for standard IMAP/POP mail servers. It also has GMail integrations.
The mail is usually downloaded into local mbox files.
Ah understood. I'm not familiar with the mbox format but from a quick search it looks like they're text files so should be easy enough to build support for. Let me test this and get back
I would be happy with incorporating Mac mail search into your tool. So you just proxy spotlight/mail search capabilities.
Understood, we'll look into this. Thank you!
Discord link is broken.
Should be fixed now - https://discord.com/invite/R63PTQ6CFh
Interestingly, both IBM and Lexmark trademarks on Selectric™ are dead or expired, and other marks don't conflict.
You likely want to file to claim this.
https://www.trademarkia.com/search/trademarks?q=SELECTRIC&co...
Immediately thought this was an odd name choice. Why bother with the namespace overlap? Pick something original.. Branding 101
I thought the same, then checked the marks, and found them expired.
If the original owners are so sure the names are dead they let the marks expire, well, "SELECT" electric is pretty good, and fewer people than ever have any exposure to the original, so why not?
Yep, suppose if you're banking on no one really knowing what it was, fairplay. But in meantime, good luck showing up in any searches for the name.
Good callout! Paperwork is in progress