Viktor.comAn AI coworker that actually does the work
Launched this week
Actx0
Memory infrastructure for AI agents.
108 followers
Memory infrastructure for AI agents.
108 followers
Your agents forget everything the moment a session ends. You stuff more context into every prompt, burn tokens on redundant history, and still ship responses that feel like amnesia with extra steps. Actx0 is the memory layer your agents are missing — a drop-in infrastructure that stores what matters, retrieves it in milliseconds, and keeps working across sessions, agents, and apps. Built for production teams who care about latency, cost, and control.
Actx0
We spent few months building Actx0 because AI agents have amnesia.
Every new session feels like a first date. To keep them coherent, developers have to constantly feed old data back into prompts; paying a massive token tax.
Actx0 fixes this. It is a managed memory infrastructure that extracts what matters and serves it back in milliseconds. No vector-store babysitting, no bloated prompts. Just true, persistent memory.
Look, we are just getting started so don’t expect a flawless, final product yet. this is day 1 of a massive roadmap, and we are shipping fast.
Paid plans are coming soon, but you can jump in and use it for free right now. Help us build the future of AI memory, try it out, and tell us what do you think!
"no vector-store babysitting" is the line that got me since that's exactly the part of our agent stack that eats the most eng time right now. We run agents on behalf of multiple customers though, so the question that matters most for us is isolation - is tenant separation enforced at the storage layer itself (separate namespaces/keys per customer) or is it something we'd have to get right ourselves in how we tag and query memories? A memory leak across customers would be a much worse bug for us than an agent just forgetting something.
Actx0
@galdayan No indeed it is isolated per workspace and you can extra isolate them with tags like per team or per agent. Same applies to messages and memories. messages of a specific user or memories are totally isolated from other users under the same workspace.
Paid plans will be able to create may workspaces. each has his own plans, members, audit ... etc
Actx0
@galdayan actually one of the features we will provide is attaching a certain knowledge to agents and be able to send Just the user query and we will give you the context whether from the rag or previous messages or memories. I believe this will save a lot of times for engineering teams building custom agents in prod.
Also we planning to support connectors as knowledge, you can connect Jira, Notion, google drive .. etc.
@clivernco the connectors piece is the part I'd want to dig into before trusting it with customer data. if someone revokes Actx0's access to a Notion page, or the page itself gets deleted at the source, does the memory layer detect that and purge whatever it already pulled in, or would an agent keep confidently surfacing that context after the source of truth is gone? for us that's a worse failure mode than isolation, since it would look correct in the moment.
Actx0
@galdayan I still working on it but i know it is complicated since we need sync data from source if something got deleted or updated. also chunking these data is not a small task.
Just checked out the site, love how intuitive the SDK is. And that it supports multiple frameworks also. So the cloud infrastructure is all managed?
Wishing you a great launch and congrats!
Actx0
@silvia_odwyer1 Yes all managed. Thanks!
Hello Inbox
@clivernco Can this work with OpenClaw?
@mikesabet You may find this interesting for the issues you mentioned last time we talked.
Actx0
@mikesabet @ismaelyws yes, you can write a custom OpenClaw tool using the Actx0 Python SDK (https://github.com/Actx0/Pctx0). a guides will be added to integrations for OpenClaw, Cursor, Claude Code ... etc) https://docs.actx0.com/agent-plugins/overview
@ismaelyws @clivernco
Brilliant. Thanks. I will try it out.
It has been quite frustrating at times.
I have been working with obsidian and grafiti and still running into a wall every morning
Cheers for this one