context and memory infrastructure for AI
Joined August 2025
- Tweets57
- Following7
- Followers269
- Likes49
RetainDB retweeted
I’m getting a lot of inbound from investors lately and idk what’s happening .
If you see my posts or you’ve heard about @retaindb and you are considering investing here’s our investor brief to save you some time
docs.google.com/document/d/1…
Check out our investor brief
docs.google.com/document/d/1…
I’m getting a lot of inbound from investors lately and idk what’s happening .
If you see my posts or you’ve heard about @retaindb and you are considering investing here’s our investor brief to save you some time
docs.google.com/document/d/1…
RetainDB retweeted
We just shipped something big at RetainDB.
Until now, AI agents had memory, but they still worked in silos. One agent processes a document, another agent has no idea it exists. Everything has to be re-uploaded, re-parsed, re-prompted. Context keeps getting lost.
We fixed that.
RetainDB now gives agents a shared workspace.
You can upload a file once, ingest it, and it becomes part of memory. Not just stored somewhere, but actually understood. It gets parsed, chunked, embedded, and turned into structured memory your agents can search and use.
So instead of forcing context into prompts, agents can just query memory and already know what’s inside your documents.
And because files live at stable URIs like rdb://files/…, they’re not temporary links or blobs that disappear. They’re first-class references agents can pass around, reuse, and build on.
One agent can process a report, another can pick it up later, extract insights, and continue the workflow. No duplication, no context loss.
(this is only available on the cloud version the oss version doesnt have this ...yet)
We just shipped a big update at retaindb .
We just shipped something big at RetainDB.
Until now, AI agents had memory, but they still worked in silos. One agent processes a document, another agent has no idea it exists. Everything has to be re-uploaded, re-parsed, re-prompted. Context keeps getting lost.
We fixed that.
RetainDB now gives agents a shared workspace.
You can upload a file once, ingest it, and it becomes part of memory. Not just stored somewhere, but actually understood. It gets parsed, chunked, embedded, and turned into structured memory your agents can search and use.
So instead of forcing context into prompts, agents can just query memory and already know what’s inside your documents.
And because files live at stable URIs like rdb://files/…, they’re not temporary links or blobs that disappear. They’re first-class references agents can pass around, reuse, and build on.
One agent can process a report, another can pick it up later, extract insights, and continue the workflow. No duplication, no context loss.
(this is only available on the cloud version the oss version doesnt have this ...yet)
RetainDB retweeted
@retaindb is now live in Hermes Agent memory update @NousResearch
Literally a one-minute setup: just run "hermes memory setup", add your API key, and your agent gets memory.
We handle storage, retrieval, and context across sessions.
Also it’s open source
RetainDB retweeted
If you have questions about the @retaindb integration in Hermes agent, happy to answer any questions
We’d be creating a discord server for all our users soon .
Hermes Agent now supports @plastic_lab's Honcho, @mem0ai, @openvikingai, @Vectorizeio's Hindsight, @retaindb, and @ByteroverDev memory systems!
Try them now with `hermes update` then `hermes memory setup`
We have rehauled our memory system to be much more maintainable and pluggable, so anyone can make their own memory system to build on top of Hermes easily and cleanly with a special class of plugin!
Which memory system is your favorite?
RetainDB retweeted
Hey @Teknium the retaindb setupin hermes was broken we made a fix pr github.com/NousResearch/herm…. please review
Hermes Agent now supports @plastic_lab's Honcho, @mem0ai, @openvikingai, @Vectorizeio's Hindsight, @retaindb, and @ByteroverDev memory systems!
Try them now with `hermes update` then `hermes memory setup`
We have rehauled our memory system to be much more maintainable and pluggable, so anyone can make their own memory system to build on top of Hermes easily and cleanly with a special class of plugin!
Which memory system is your favorite?
Hermes agent now supports retaindb
Excited to see who tried us out
Hermes Agent now supports @plastic_lab's Honcho, @mem0ai, @openvikingai, @Vectorizeio's Hindsight, @retaindb, and @ByteroverDev memory systems!
Try them now with `hermes update` then `hermes memory setup`
We have rehauled our memory system to be much more maintainable and pluggable, so anyone can make their own memory system to build on top of Hermes easily and cleanly with a special class of plugin!
Which memory system is your favorite?
We built an mcp that adds context to your coding agent
Add it to your coding assistant today!👇
Your coding agent always starts with zero context.
No memory of last week’s decisions.
No searchable knowledge of your repos, docs, or videos.
We just shipped RetainDB MCP the persistent context layer for any MCP-compatible agent.
Index your GitHub repos, Notion, PDFs, YouTube videos (transcribed & searchable), Confluence, Slack & more.
Then just ask and your agent instantly pulls the right context + codebase with smart semantic search.
One command to get started:
npx retaindb-wizard
Works with Cursor, Claude Code, Claude Desktop, Windsurf & more.
Never explain your project from scratch again.
retaindb.com/mcp
RetainDB retweeted
it's an uphill battle I am fighting here tbh , it's hard to complete with vc backed startups that do the same thing as i am doing.
Lots of memory based startups popping up these days so if you’re good at gtm / distribution and would like to co found the best memory solution for AI in the market please DM
This is not me complaining btw but I feel like I need to focus on product development rather than all , before I get spaced out .
Startup is @retaindb btw
Big win for open-source agents!
Hermes Agent now has native, durable memory powered by @RetainDB
true cross-session recall, persistent fact storage, and skills that actually evolve over time.
No more forgetting. No more manual hacks. Memory is now a first-class citizen
@NousResearch Hermes agent now has durable memory
We added native RetainDB memory to Hermes.
That gives Hermes built-in cross-session memory with:
• native CLI setup
• recall across sessions
• durable session ingestion
• native memory tools
The important part is that memory now lives as a first-class capability inside Hermes, not just transcript replay or manual note-passing between chats.
PR:
github.com/NousResearch/herm…
@Teknium @NousResearch
Me after I discovered I could literally add memory to my agent in 3 lines
We shipped a LangChain + LangGraph integration for RetainDB.
So now your agent can retrieve relevant user context before replying, then store new turns back into memory for future sessions.
Not chat history glued on top.
Actual persistent memory.
Repo: github.com/RetainDB/retaindb…
RetainDB retweeted
We built RetainDB because AI with no memory is just an expensive search bar.
Every AI app today handles context the same way: stuff it in the prompt, hope it fits, start over next session.
Stateless, sure. Useful? Almost never.
Your LLM doesn’t know what you decided last Tuesday.
It doesn’t know the project changed. It doesn’t know you already tried that approach and it failed.
A founder I know spent 3 hours re-explaining his codebase to his AI assistant. Every. Single. Day.
Once you’re building anything non-trivial, this compounds into pure friction.
AI accuracy goes to shit the moment context resets.
We built @retaindb for exactly this.
RetainDB gives your AI persistent memory + grounded retrieval — so it knows what was decided, what changed, and why, across every session.
0% hallucination rate on our benchmark. 10,000+ turns retained without degradation.
Your AI should remember. Not start over.
More below
RetainDB retweeted
@usewhisper is all you need .
npm I -g @usewhisper/sdk
