Pinned Tweet
ChatAndBuild's Living Brain ranks No. 1 on LoCoMo, with our evals reporting 92.2% judged accuracy, compared to a full-context, model-only baseline run of 72.3%.
Living Brain is model-independent memory layer for AI agents. It turns documents, meeting notes and conversations into structured, evolving memory so agents can build on what you already know across AI platforms.
Integrate @livingbrain into your applications and agent systems through its API, MCP, SDK and CLI on livingbrain.com.
Living Brain is model-independent memory layer for AI agents. It turns documents, meeting notes and conversations into structured, evolving memory so agents can build on what you already know across AI platforms.
Integrate @livingbrain into your applications and agent systems through its API, MCP, SDK and CLI on livingbrain.com.
We decoupled memory from the model.
Living Brain is one persistent brain for ChatGPT, Claude, Cursor, Codex and your own agents. Write in one, retrieve in any.
92.2% on LoCoMo. 464ms median retrieval. Up to 22.6x fewer tokens per question.
LLMs have parameters. Living Brain has plasticity.
14
10
48
2,461
Switching to @ChatGPT or Codex from another coding agent used to mean starting over.
With a @livingbrain, your preferences, your ways of working and your past conversations come with you.
OpenAI's Sun Weiran explains 👇
With a @livingbrain, your preferences, your ways of working and your past conversations come with you.
OpenAI's Sun Weiran explains 👇
"I gave Codex a Living Brain."
OpenAI's Sun Weiran on carrying your preferences and past conversations into Codex and ChatGPT, so every AI app feels "truly unified and personalized".
Full interview 👇
1
0
2
112
ChatAndBuild retweeted
What a week. We launched @livingbrain.
5.5M views in 72 hours.
Trending X story with 2,311 posts created about us.
#2 on @ProductHunt (even ahead of @OpenAI)
Shipped our official @ChatGPT plugin
Showed how easy it was to go from Claude Code to Codex and more with just @livingbrain.
And we’re just getting started. LFG.
5.5M views in 72 hours.
Trending X story with 2,311 posts created about us.
#2 on @ProductHunt (even ahead of @OpenAI)
Shipped our official @ChatGPT plugin
Showed how easy it was to go from Claude Code to Codex and more with just @livingbrain.
And we’re just getting started. LFG.
We decoupled memory from the model.
Living Brain is one persistent brain for ChatGPT, Claude, Cursor, Codex and your own agents. Write in one, retrieve in any.
92.2% on LoCoMo. 464ms median retrieval. Up to 22.6x fewer tokens per question.
LLMs have parameters. Living Brain has plasticity.
25
17
114
198,187
#2 on Product Hunt, and featured in today's Product Hunt newsletter. 🧠
Proud of this team. @livingbrain is one persistent memory across ChatGPT, Claude, Cursor and Codex.
This is the foundation for everything we build next. 👇
Proud of this team. @livingbrain is one persistent memory across ChatGPT, Claude, Cursor and Codex.
This is the foundation for everything we build next. 👇
Living Brain is #2 on @ProductHunt 🧠
And it just got featured in the Product Hunt newsletter, in front of 350k+ people.
Come say hi 👇 producthunt.com/products/chata…
1
0
6
221
OpenAI shipped 20 major announcements in a single day.
Sun Weiran, Applied AI Engineer Lead at @OpenAI, on why agents that keep running need a shared memory.
Chapters:
0:00 Intro
1:21 OpenAI DevDay announcements and releases
3:16 Key industry trends and model advances
9:16 Living Brain
11:00 Developer experience and ecosystem tools
17:51 Conclusion and final takeaways
Set up Living Brain and try it yourself 👇
livingbrain.com
0
2
12
324
ChatAndBuild retweeted
switching between AI models is easy. bringing all your context with you is the annoying part.
@livingbrain is building a shared memory layer for ChatGPT, Claude, Codex and other agents, so you don't have to keep explaining the same things.
it figures out what's worth es what matters instead of feeding in your entire history.
their internal tests show up to 22.6x fewer tokens per question, with 464ms median retrieval.
pretty useful if you work across multiple AI tools.
@livingbrain is building a shared memory layer for ChatGPT, Claude, Codex and other agents, so you don't have to keep explaining the same things.
it figures out what's worth es what matters instead of feeding in your entire history.
their internal tests show up to 22.6x fewer tokens per question, with 464ms median retrieval.
pretty useful if you work across multiple AI tools.
We decoupled memory from the model.
Living Brain is one persistent brain for ChatGPT, Claude, Cursor, Codex and your own agents. Write in one, retrieve in any.
92.2% on LoCoMo. 464ms median retrieval. Up to 22.6x fewer tokens per question.
LLMs have parameters. Living Brain has plasticity.
4
6
39
1,363
ChatAndBuild retweeted
You know how you use one AI, teach it about yourself, your work and your preferences?
Now imagine switching to a completely different AI and it already knows all of that.
Your projects, previous conversations, important details and instructions are still there.
That’s what @LivingBrain makes possible.
It gives different AI models and agents access to the same evolving memory.
No repeating yourself. No starting from scratch. Just one shared memory across your AI tools.
Imagine how much time that could save. x.com/ladyxtel/statu…
Now imagine switching to a completely different AI and it already knows all of that.
Your projects, previous conversations, important details and instructions are still there.
That’s what @LivingBrain makes possible.
It gives different AI models and agents access to the same evolving memory.
No repeating yourself. No starting from scratch. Just one shared memory across your AI tools.
Imagine how much time that could save. x.com/ladyxtel/statu…
We decoupled memory from the model.
Living Brain is one persistent brain for ChatGPT, Claude, Cursor, Codex and your own agents. Write in one, retrieve in any.
92.2% on LoCoMo. 464ms median retrieval. Up to 22.6x fewer tokens per question.
LLMs have parameters. Living Brain has plasticity.
7
12
53
27,236
ChatAndBuild retweeted
I’ve not tested this, but it has peaked my interest enough that I’ll try to spend my flight over to Adobe Max playing around with this.
Shared memory, across any agentic provider essentially is the most painful thing at the moment.
I hope this could be a solution
Shared memory, across any agentic provider essentially is the most painful thing at the moment.
I hope this could be a solution
We decoupled memory from the model.
Living Brain is one persistent brain for ChatGPT, Claude, Cursor, Codex and your own agents. Write in one, retrieve in any.
92.2% on LoCoMo. 464ms median retrieval. Up to 22.6x fewer tokens per question.
LLMs have parameters. Living Brain has plasticity.
50
9
63
9,377
ChatAndBuild retweeted
Give your @ChatGPT a @livingbrain.
It’s live in Plugins:
1. Search “Living Brain” and install
2. New chat: “Save my memories to Living Brain.”
3. Then: “Show me my Living Brain.”
Those memories now work in Claude, Cursor and Codex too. Install once, use everywhere.
It’s live in Plugins:
1. Search “Living Brain” and install
2. New chat: “Save my memories to Living Brain.”
3. Then: “Show me my Living Brain.”
Those memories now work in Claude, Cursor and Codex too. Install once, use everywhere.
We decoupled memory from the model.
Living Brain is one persistent brain for ChatGPT, Claude, Cursor, Codex and your own agents. Write in one, retrieve in any.
92.2% on LoCoMo. 464ms median retrieval. Up to 22.6x fewer tokens per question.
LLMs have parameters. Living Brain has plasticity.
46
28
178
829,895
ChatAndBuild retweeted
@livingbrain is model-independent memory layer for AI agents. It turns documents, meeting notes and conversations into structured, evolving memory so agents can build on what you already know across AI platforms.
Read full here 👇 x.com/i/article/2118…
Read full here 👇 x.com/i/article/2118…
9
9
43
2,148
Living Brain ranks No. 1 on LoCoMo with 92.2% judged accuracy across 1,540 questions, using close to six times fewer tokens than full context.
1
0
3
24
Yesterday: 1.8m views and trending on X. 🧠
Today: we're live on @ProductHunt.
People are tired of re-explaining their work to every AI tool. @livingbrain gives them one memory across @ChatGPT, @claudeai, @cursor_ai and Codex.
Come say hi 👇
producthunt.com/products/chata…
Today: we're live on @ProductHunt.
People are tired of re-explaining their work to every AI tool. @livingbrain gives them one memory across @ChatGPT, @claudeai, @cursor_ai and Codex.
Come say hi 👇
producthunt.com/products/chata…
25
6
38
3,013
ChatAndBuild retweeted
Switching AI tools shouldn’t mean starting from scratch every time 😭
@livingbrain lets ChatGPT, Claude, Cursor and Codex share the same project context.
One memory across your tools. This is so needed.
@livingbrain lets ChatGPT, Claude, Cursor and Codex share the same project context.
One memory across your tools. This is so needed.
We decoupled memory from the model.
Living Brain is one persistent brain for ChatGPT, Claude, Cursor, Codex and your own agents. Write in one, retrieve in any.
92.2% on LoCoMo. 464ms median retrieval. Up to 22.6x fewer tokens per question.
LLMs have parameters. Living Brain has plasticity.
4
4
22
1,557
ChatAndBuild retweeted
last year i said the real bottleneck for AI agents wasn't intelligence; it was memory
we've spent so much time scaling model capabilities, context windows and inference-time compute while agent memory still largely relies on messy Markdown files, vector databases and repeated ous how LivingBrain architecture handles this problem by making memory persistent, adaptive and independent of the model or agent harness
we've spent so much time scaling model capabilities, context windows and inference-time compute while agent memory still largely relies on messy Markdown files, vector databases and repeated ous how LivingBrain architecture handles this problem by making memory persistent, adaptive and independent of the model or agent harness
We decoupled memory from the model.
Living Brain is one persistent brain for ChatGPT, Claude, Cursor, Codex and your own agents. Write in one, retrieve in any.
92.2% on LoCoMo. 464ms median retrieval. Up to 22.6x fewer tokens per question.
LLMs have parameters. Living Brain has plasticity.
30
13
75
4,044
ChatAndBuild retweeted
Major breakthrough for agent memory: @livingbrain says it hit 92.2% judged accuracy on LoCoMo across 1,540 questions, with 464ms median retrieval.
It also reports health scores for every brain, freshness, strain and readiness, so you see stale sources before your agent uses them.
These are their own numbers. If they hold up in independent tests, memory stops being the weak spot in every agent.
It also reports health scores for every brain, freshness, strain and readiness, so you see stale sources before your agent uses them.
These are their own numbers. If they hold up in independent tests, memory stops being the weak spot in every agent.
We decoupled memory from the model.
Living Brain is one persistent brain for ChatGPT, Claude, Cursor, Codex and your own agents. Write in one, retrieve in any.
92.2% on LoCoMo. 464ms median retrieval. Up to 22.6x fewer tokens per question.
LLMs have parameters. Living Brain has plasticity.
17
5
90
8,659
This is the milestone we've been building toward. 🧠
1.4m views and trending on X in under 12 hours. We set out to build AI that carries your work forward, not AI that forgets you.
Thank you for showing up for it @livingbrain
1.4m views and trending on X in under 12 hours. We set out to build AI that carries your work forward, not AI that forgets you.
Thank you for showing up for it @livingbrain
Gosh what a morning. Thank you everyone, completely overwhelmed by the response.
@livingbrain did 1.4m views in less than 12 hours, and now we are trending on X!
Just the beginning. x.com/i/trending/211…
10
3
20
575
ChatAndBuild retweeted
Why "Living Brain"?
Brains keep what matters, link it to what they already know, and recall it when it's relevant. That's neuroplasticity, and it's what we designed around.
4 scoring policies decide every memory:
① Worthiness: is it worth keeping?
② Placement: where does it at does it connect to?
④ Re-entry: does it matter for this task?
The goal: the most useful information per inference token.
Brains keep what matters, link it to what they already know, and recall it when it's relevant. That's neuroplasticity, and it's what we designed around.
4 scoring policies decide every memory:
① Worthiness: is it worth keeping?
② Placement: where does it at does it connect to?
④ Re-entry: does it matter for this task?
The goal: the most useful information per inference token.
1
15
144
33,738
ChatAndBuild retweeted
Your agent keeps its brain when you change its harness.
Connect OpenClaw, Hermes or your own agent through the API, SDK or MCP. Switch frameworks and the accumulated context comes with you.
The brain lives outside the runtime.
Connect OpenClaw, Hermes or your own agent through the API, SDK or MCP. Switch frameworks and the accumulated context comes with you.
The brain lives outside the runtime.
8
19
261
64,225
Byte has been rehearsing all week. Brainy remembers every line.
The show starts tonight. 🦖🧠
#LivingBrain
The show starts tonight. 🦖🧠
#LivingBrain
7
6
19
2,141