Launched this week
Unabyss
MCP-native self-updating context layer for your AI
1.6K followers
MCP-native self-updating context layer for your AI
1.6K followers
Set it up once and never re-explain yourself to AI again. Connect the apps you use daily - Unabyss will extract, structure, and update your context automatically. Share it with any AI tool via MCP, with granular control over what each tool can see.
Unabyss
@vihaan_pande thanks! At the moment we dont have those integrations but we are working on it ;)
Unabyss
@vihaan_pande yes there is, you can upload your own files in the Sources page ;)
LayerProof
This sounds helpful, just a quick question about the control part: If I connect to my apps, can I tell Unabyss to only look at specific folders, or does it just scan everything it finds?
Unabyss
@creativewjordyn it depends on the app. In Slack, Notion, Calendar, and GitHub, you can choose exactly what to sync. In other apps, we sync all available data, filter it, and segment it.
Curious how context staleness is handled — if I update a doc in a connected app mid-conversation, does the MCP layer reflect that in real-time or on a sync interval?
Unabyss
@hirogure great question! you can set up the sync interval for each app connected to Unabyss (see the screenshot).
In the situation you described, context might be missing the update, but after 5min, it'll be imported & added to the vault.
do you see specific use-cases where the real-time sync would be essential?
when linkedin says one thing and notion says another, who wins? the conflict resolution between sources is where this gets tricky and most context tools just pick the latest update which isn't always right
Unabyss
@tina_chhabra we have a few soultuons for that. One starts when you onboard. Unabyss creates an identity summary that you can tweak or add to it. This is a baseline of who you are. Then, whenever a new date comes up and differs from what we already have, the agent will ask you to decide what's true and what's not.
Congratulations on the launch! This is a great product and much needed.
One question I had: I have setup memory.md files in Claude which is basically a memory of the interactions I have had with Claude for it to remember the context of the project. Can Unabyss understand from those files about "how" the engineer usually proceeds with a problem?
Unabyss
@aiswarya_s great question!
Yes, you can import these files to Unabyss and they will become a natural part of your context. Plus, from the moment you connect with MCP, Unabyss will know your conversations, which will help it improve - compounding effect at its best :)
What else can I share to help? :)
This sounds great and I was actually waiting for something like this. My main question: is there a way to handle projects or specific topics separated? For instance when working on several codebases / products. My issues is always the amount of information I need to manage across projects while keeping an overview. Currently it feels like one huge mixed salad.
Unabyss
@arkadiy_kreslov1 Yes, you can do it in many ways. One is to generate separate exports that will have all the defining info about each project, then you can use them as is (they will update along the way) or use them as a starting point for agents to retrieve more details connected to the project. But that's a more manual way to do it.
By design, Unabyss segments data into different "buckets," so new information about a certain project will be automatically connected to it. You just need to ask about it ;)
Love the idea of a centralized context vault! Repeating the same brand guidelines and operational context to different AI tools is a massive friction point. Quick technical question: when sharing data via MCP, how does Unabyss filter what's actually relevant to the user's current prompt so it doesn't just nuke the LLM's context window and skyrocket token costs?
Unabyss
@andika_fadhilah great question, Andika! There's specific agent that decides what kind of information can be useful for this specific prompt :) So we don't nuke the context window - it's a precise & accurate context ingest