There is a lot of competition in this space, both commercial companies and open source. It seems like the biggest advantage for using one of these and not Claude or Codex is that you can use multiple models. Otherwise, those two have all the same features, or probably will in the next week.
Are you able to use multiple models? Or I guess the first question is, what model(s) are you using? At this point, many enterprises care so that should be front and center, especially if you're using models hosted in China, because a lot of non-Chinese companies care about that.
You mention some things about unique ways to store memory. Do you have any data or use cases on how that improves performance? If you do, you should get those on your website too.
I'm not sure you have a very strong answer to "Why not just run Claude or Codex on my laptop?" Sure, you provide compute, but so do they, and most enterprises and even regular people are getting desktops/droplets to run their agents now because of this problem. This is kind of solved already.
Also there is a small nit with your website. The graphic next to "Proactively gets things done for you" gets bigger and smaller, so if you're trying to read anything below that, it keeps jumping up and down.
Currently we don't allow for multiple model selection, but it is on the roadmap. We do have a cli, which allows you to use the wiki layer and the tools in whatever harness you prefer. Further, we want to allow users to also bring their own keys/subscription.
We do not have formal benchmarks to measure improvement in performance. A big reason is that most benchmarks are very "atomic facts memory" focused. They'd give a huge passage, conversations, and see which memory system could surface those facts. How we're different is agents could be simply given the markdown files, and they can pretty much get the perfect context themselves. Second, we are banking more on the fact that having this layer enables new capabilities of the models
> Currently we don't allow for multiple model selection, but it is on the roadmap.
Cool! What model does it use now? Enterprises will definitely want to know.
And the other big question: Why not just use Claude or Codex (via CLI on a desktop or droplet)?
Its GPT-5.6-terra under the hood. Also we want to control the harness. Features like proactivity/long tasks are not possible the way Claude/Codex work today. We do plan on releasing our desktop app, which would have these things inbuilt.
Hi ! Are you guys looking for interns or smth right now , this looks like something I would want to work on xD
> that knows everything about your company
Given that I don't know everything about my 4-person company, and that I never knew everything when it was a 1-person company (guess which person is that), this sound like https://x.com/mitchellh/status/2044529496217555333 (apologies for twitter link)
why "apologies for twitter link"?
Don’t have time to dig deep into it this afternoon, but want to comment to come back but also to suggest an expansion past your stated “one personal and one company” wiki.
I lead two different functional areas within tech and work in a larger tech platform for the parent company of a couple dozen operating companies.
All that to say “I would almost surely be better served by a hierarchy or general graph of wikis rather than just ‘me’ or ‘all 15K employees’.”
Even a 50 person company might benefit from “me”, “my team”, “my department”, and “whole company”.
I guess you could say, you weren't cimpress-ed
How do you manage permissions in the wiki? Does it require all content being available to everyone in the org or does it do some sort of vectorizing that is unique to every single user based on their perms? Whenever we've tried to build something similar, permissions is usually the first headache and is often why something like Microsoft's WorkIQ is appealing (if you're an O365 org).
The current permission implementation is binary. It is either personal to you, or is shared with everyone in the company. Having more granular permissions is something that has been requested often, and it is on our roadmap. I'm curious, how does Microsoft WorkIQ handle these permissions?
If AI has achieved anything it's a dramatic increase in documentation quality for the sake of training or instructions.
I'd argue that if documentation on how to operate your business can be used to feed an agent, your problem was never the ability to run operations and the ROI will amount to the ability to ask instead of doing. I don't know how much this is gonna go for, but you could probably have one of your existing staff just do basic documentation lookups too.
The documentation is not just for lookup queries. Having this layer betters the quality of responses, but also enables a lot of features we have in Almanac/plant to add. I mentioned in the post how this layer enables proactive agents and long horizon tasks. I just came across a Google paper yesterday (https://arxiv.org/html/2608.27454v1) that use a similar wiki knowledge base to self improve agent skills. I think these are just scratching the surface of ideas of what's possible with this layer
The source-retention detail seems more important than the model list: if an email is compiled into a wiki and later deleted, or access is revoked, how quickly does that fact disappear from the personal or company wiki? Is the wiki rebuilt from current source state, or are removals handled as a separate event?
Wiki updates happen once a day through. That is when the entire day's context is updated in the wiki. That is when the fact disappearance would take place
How are you different from the following cos?
OpenTag
Turnstone
Memory Store
Hyper
Hyperspell
Wato
Savant
Shepherd
Glen
OpenTag
Janet AI
screenpipe
Cerenovus
jo
Bond
Airweave
Mem0
Theta
Nessie
Mosaic
Onyx
Slite
Dashworks
Circleback
Stamp
Tasklet
Within
Khoj
There is truly a lot of competition in the space. The reason we're confident we stand out is because of our wiki layer. We have seen both the benefits of having a detailed compiled wiki, and more importantly, actually using this layer to then power capabilities. Two such examples are written in the post as proactive agents and the handling long horizon tasks. I believe this is still just scratching the surface. Having the perfect context layer makes every single task you do with agents much moe powerful
Wishing you guys the best, but this isn't a very compelling answer. Would be helpful to provide more concrete or quantifiable examples. In the early stages when pitching to users, it's preferable to state exactly what can be done now.
7-day trial seems kind of narrow for a product that needs to "build a memory", no?
It of course gets more magical the longer you use it, but we've felt it starts providing utility from first few days itself. It could identify things that are going on immediately in your life and help there
For us, these two things have not been very different. We do make both, a personal and a company wide wiki, so Almanac is useful in both regards
Microsoft WorkIQ but for companies on a non-MS stack
Just checked out WorkIQ, the promise is quite similar. Spend compute and time upfront developing the knowledge base, which will later result in much better answers
The website tells me absolutely nothing about what makes this new or unique
I would agree that there are a lot of people in this space, both products and roll your own. We are doing our own stuff here in our company right now, and could be interested. But for me, I didn't feel like i know enough to actually assess and take the next steps. Good luck!
I'd love to hear how you're approaching this problem, and what would you like to see on the website to convince you to try out Almanac
Given it has "logins" to tools without APIs (browser-based access), how do they handle 2FA, session persistence, and ToS violations for services that don't allow bot access?
Our browser is interactive, in the sense that one could click around the browser, as well as the agent. Usually for 2FA logins, Almanac would hand it over to the human, who'd complete login, and then continue.
We use Browserbase for the browser and session persistence is inbuilt.
For ToS violations, this is a limitation. Accessing sites that don't allow for bot access could result in account ban/suspension. Simply using Browserbase under the hood does not remove this risk.
Really simple and well designed website for the brand identity. Logo, background, UI and FAQ all flows well. Feel like I trust the project without knowing much about it or having the usual “trusted by” section. Keep it up!
Only nitpick maybe remove the red box next to the Pricing label on the pricing page. Looks like the Polish flag on first glance and I immediately thought “does it show me pricing from Poland”?
Congrats on the launch! Really useful tool!
Congratulations on the launch! The demo in the video you linked looks awesome (though your video resolution is a bit low, some of the text is hard to read on screen, I suggest you record in higher resolution). I think you made a really good choice to have your connector provider handle OAuth credentials, since it is really hard to get security right on a multitenant app like yours, where some of your agents are doing work on customers' behalf and others are improving your product itself :). The design of your app interface also looks really great!
Thanks! I saw the video again and can't look beyond the low resolution now lol. We knew from day 1 that trusting agents is going to be the most difficult part of our product, so we take privacy and security with the utmost seriousness