A Second Brain That Helps My AI Agents Share Context
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: A Second Brain That Helps My AI Agents Share Context on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get hardware and tech essentials delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

Thorsten Meyer says he built a shared “Second Brain” so the AI agents he runs can retrieve and add to one long-term memory. The system uses Markdown files, a one-line-per-page index and Git rather than a vector database or cloud service; Meyer reports 1,278 pages and a 97.4% top-one recall score as of November 21, 2026.

Thorsten Meyer has described a shared long-term memory system for the AI agents he runs, allowing each agent to look up and add work-related context instead of starting every session without it. In an article published November 21, 2026, Meyer said the system contains 1,278 pages and reported a 97.4% top-one recall score, a measure he defines as how often its first selected page is the right one.

Meyer calls the system his Second Brain. Its records are plain Markdown files kept in Git, with each page grouping related information, such as material about a project or a decision. A single file, INDEX.md, lists each memory on one line. Meyer says there is no vector database or cloud service in this setup, and that the files can be opened in an ordinary editor.

When an agent needs information, a script named recall.py scores the index entries first, then opens the best-matching page and returns the section it identifies as relevant, along with the source. Meyer says this design lets him check where an answer came from. A separate script, remember.py, writes a new page, its index entry and a log entry together, a process intended to keep the index aligned with the stored notes.

As of November 21, Meyer reports 1,278 pages, 1,761 links and 1,363 recalls. He also gives the system a 97.4% top-one recall score. These are figures reported by Meyer; the article does not describe an independent evaluation or provide details such as the test set, measurement period or how “right” was judged.

At a glance
announcementWhen: Described November 21, 2026
The developmentThorsten Meyer published a description of a shared, file-based memory system that lets his AI agents retrieve and add work context across sessions.
Crypto market snapshot
Fear & Greed Index
12/100 — Extreme Fear
Bitcoin BTC$60,316▼ 2.8%
Ethereum ETH$1,693▼ 3.5%
Tether USDT$0.9988▼ 0.0%
BNB BNB$547.44▼ 4.1%
USDC USDC$0.9994▼ 0.0%
XRP XRP$0.9309▼ 2.0%
Solana SOL$70.65▼ 1.1%
TRON TRX$0.3055▲ 0.0%
Live data · CoinGecko · alternative.me (24h change)

A Shared Memory Across Agent Sessions

The system addresses a practical limitation of session-based AI work: information an agent can see during one session may not be available when another session begins. A shared store gives Meyer’s agents a common place to retrieve prior decisions, project terminology and lessons from earlier work. That could reduce repeated explanations and make it easier for different agents to work from the same records.

The approach also puts inspectability and portability ahead of an opaque, tool-specific memory store. Because the records are text files under Git, a person can read them, review changes and potentially move them between tools. But Meyer describes a personal system, not a validated team product. Its usefulness depends on accurate notes, well-maintained index entries and the recall script selecting relevant material; the reported score alone does not establish how it performs for other users or tasks.

Products Worth Considering

Apple MacBook Air 15-inch Laptop with M5 Chip, 16GB Memory, 512GB SSD, Midnight

Apple MacBook Air 15-inch Laptop with M5 Chip, 16GB Memory, 512GB SSD, Midnight

LG UltraGear 34-Inch Curved QHD Gaming Monitor, 160Hz, FreeSync Premium

LG UltraGear 34-Inch Curved QHD Gaming Monitor, 160Hz, FreeSync Premium

Apple MacBook Air 13-inch with M5 Chip, 16GB Memory, 512GB SSD, Midnight

Apple MacBook Air 13-inch with M5 Chip, 16GB Memory, 512GB SSD, Midnight

As an affiliate, we earn on qualifying purchases.

From Session Context to Stored Notes

AI agents generally work from information made available in their current context window. Meyer says that when a session ends, its working context is gone, leaving him to restate information in later sessions. He says some tools offer memory features, but he wanted one shared store that he could read, edit and audit himself, rather than separate memories tied to individual tools.

The design separates retrieval from storage. The index is kept short enough to scan without opening every page, while each page contains a cluster of related facts. Git provides a change history, and the log records additions in sequence. Meyer also describes nine views for monitoring and using the system, including a map of connections, a daily brief, a triage process for session captures, a change chronicle and a live view of agent activity. Those views sit on top of the file-and-index structure rather than replacing it.

“I got tired of being their memory. So I built them one.”

— Thorsten Meyer, describing the reason for building the system

How the Recall Score Was Measured

The published account does not specify how Meyer calculated the 97.4% top-one score, including the number and type of test queries, the evaluation period or who judged whether a page was correct. The figure should be treated as Meyer’s own measure, not as an independently verified benchmark or a guarantee of accuracy on other systems.

Other details also remain open. Meyer does not provide a full account of access controls, privacy protections, backup and recovery, or how conflicting or outdated memories are handled beyond the described triage process. He says the setup is built for one owner rather than a team, so its suitability for shared organizational use has not been established.

From Personal Tool to Team Use

Meyer’s article presents the Second Brain as an existing personal system and describes its current components and reported size; it does not announce a release date, public product launch or independent test. The next practical questions are whether the design can be reproduced by other users and how it would handle multiple people editing or relying on the same memory.

For now, the account’s next developments are tied to the system’s ongoing use: maintaining the index and records, reviewing session captures, and tracking how agents retrieve and add information. Any broader assessment would need more detail on evaluation methods, reliability and safeguards, particularly before treating the reported results as applicable beyond Meyer’s own workflow.

Key Questions

What is Thorsten Meyer’s Second Brain?

It is a personal knowledge base and long-term memory that Meyer says he shares with the AI agents he runs. Agents can retrieve information from it and add new records.

How does the system store information?

It stores related facts in Markdown files kept in Git. A central INDEX.md file lists the pages, and Meyer says the setup does not use a vector database or cloud service.

What does the 97.4% recall figure mean?

Meyer defines top-one recall as how often the first page selected by the system is the right one. His article reports 97.4% but does not provide enough testing detail for independent verification.

Is the system available for teams or the public?

The source describes Meyer’s own setup and says it is built for one owner, not a team. It does not announce a public release or establish that the system is available as a product.

What remains unknown about the system?

The article does not detail the recall evaluation, access controls, backup arrangements or performance in other users’ workflows. Those questions limit what can be concluded from Meyer’s reported figures.

Source: ThorstenMeyerAI.com

Nothing in this article is financial or investment advice. Cryptocurrency and precious-metal investments carry significant risk — do your own research and consider a licensed advisor.
You May Also Like

DeepSeek Sends U.S. Stock Markets Tumbling—Here’s Why Investors Are Panicking

Beneath the surface of DeepSeek’s AI launch lies a brewing storm in the stock market—what could this mean for tech and national security?

10 Best Gaming Motherboards In 2026

Explore the 10 best gaming motherboards of 2026, featuring platform choices, key features, and what makes each ideal for different gamers and builders.

Cloud’s Hidden Memory Bill

Memory shortages are quietly increasing cloud costs, with prices rising up to 15% for some services, impacting businesses relying on cloud infrastructure.

10 AI Milestones That Will Mark 2026

A forecast of ten major AI developments set to occur by 2026, highlighting technological advances and their potential impact across industries.