Central Brainby NeuroAIgent

Your AI forgets everything. Central Brain doesn't.

Central Brain is a local-first AI memory layer for Claude Desktop, Claude Code, GitHub Copilot and the ChatGPT desktop app. In the default project workflow, your decisions, facts, open questions and project files live on storage you choose and stay in files you can open. Keygen receives licensing and device data for activation and verification, not documents; connected AI tools receive the context they retrieve, and optional cloud connections send selected data only when you enable them.

Get Central Brain, $12/month See how the memory works

Windows and Apple silicon Mac
Local MiniLM embeddings
Flat per-license pricing
Memory in plain files

memory/brief.mdcheckpoint · 2 min ago

## Decisions
- Pricing stays flat per license.      09-15
- Embeddings stay local. Cloud opt-in. 09-14

## Facts
- Meridian renews 2026-11-01. 30 day notice.
- Sarah owns the security questionnaire.

## Open questions
- pgvector for the Team tier at launch?

## Todos
- [x] Send revised SOW to Meridian
- [ ] Re-index /contracts after Q3 changes

> memory_recall  new Claude Desktop session
>   2 decisions, 2 facts, 1 question,
>   1 todo recalled into this session.

Works with Claude Desktop
Claude Code
GitHub Copilot
ChatGPT desktop: Chat, Work and Codex
Over MCP, source files read-only

01The problem

Every session starts from zero. Or worse, from a summary a model wrote about you.

Built-in memory features keep a short, auto-rewritten note inside one vendor's app. Long sessions get compacted and the middle of your work disappears. Central Brain fixes the memory problem where it actually lives: on your disk, in files, across the supported tools you connect.

Compaction erases the middle of your work.
When a long session is compacted, whatever the model kept is what it guessed mattered. With Central Brain, connected assistants checkpoint decisions, facts, open questions, todos and summaries as you go, and memory_recall hands them back in the next session.
Vendor memory is a retention feature.
ChatGPT and Claude memory live in one account and work inside one app. ChatGPT's data export does not include memory entries, and since June 2026 its memory is led by a summary the model rewrites on its own. Central Brain memory is a folder you can open, copy, back up and read.
The same files, uploaded again and again.
Every new chat, the same PDFs and spreadsheets. Central Brain prepares a project once, indexes it locally, and serves it to connected tools with the source path, page number and chunk attached where available.
Tools that shut down take your memory with them.
Rewind shut down in December 2025. Limitless was acquired by Meta. Khoj Cloud sunset in April 2026. A local folder cannot be discontinued, acquired or rate limited.

02What it remembers

Five kinds of notes and a project brief. One place for every connected tool.

Central Brain keeps a per-project memory that connected AI tools can read before they answer and add to as they work. It is deliberately simple, because memory you cannot audit is memory you cannot trust.

Decisions. What you chose and when, so nothing gets re-argued next week.
Facts. Dates, names, numbers and constraints that should carry into later sessions.
Open questions. The things still unresolved, so the next session starts there.
Todos. Carried forward until they are done.
Summaries. Short recaps of where the work stands, saved like any other note.
Project brief. The compact context a new session can start from. Compaction replaces it with an updated summary.

What an AI memory layer is, and why files beat vector-only stores

MCPTools exposed to your AI

central_brain_pingstatus

Confirms the server is running and reports the project root folder.

memory_recallmemory

Searches decisions, facts, open questions, todos and the brief before the model answers.

memory_checkpointmemory

Appends one note at a time (a decision, fact, question, todo or summary) as the work happens.

memory_compactmemory

Replaces the project brief with an updated summary and marks older notes as compacted, without deleting them.

semantic_searchsearch

Vector search over prepared files with the local MiniLM model. Results carry path, page and chunk where available.

search_textsearch

Exact text search when you know the phrase and want zero guessing.

read_filefiles

Reads a file under the project root, read-only.

list_filesfiles

Lists what the project contains so the model can navigate instead of guessing.

03How it works

Five steps. Then the memory compounds.

Point Central Brain at a folder. It prepares and indexes the files locally, opens a project-specific MCP server, and connected tools get search plus memory. Refresh only touches what changed.

Choose
Pick the source folders and where prepared output should live. Those project files stay on storage you choose in the default workflow.
Prepare
PDF, DOCX, TXT, Markdown, CSV and XLSX become AI-ready text. Optional OCR, where the build includes it, recovers scans and photos.
Index
Chunks are embedded with the bundled 384-dimension MiniLM model into a local vector index. No cloud call required.
Connect
Add the project's MCP server to Claude Desktop, Claude Code, GitHub Copilot or the ChatGPT desktop app. Memory and search appear as tools.
Refresh
New, changed, removed and failed files are processed on their own. Interrupted work resumes.

The full walkthrough MCP memory server guide

04Files you can open

Local-first is a design choice, not a slogan.

In the default workflow, your source files, prepared output, project state, local vectors and memory stay on storage you choose. Keygen receives licensing and device data, not documents. Cloud embeddings, Pinecone, Postgres and Redis are optional connections you turn on, and they send selected data only when enabled.

Local credential storage for optional connections; protection depends on the operating system and installed release.
Read-only file tools: MCP tools can read and search your files, not rewrite them. Memory tools append notes and update the project brief.
Project-by-project organization and permissions.
Revocable same-Wi-Fi device trust, with sharing off by default.
Runs air-gapped: preparing files, indexing, search and memory need no network connection. The license is re-checked every 14 days, and Team and Enterprise can get a build with no online check. Cloud-hosted AI clients and optional cloud connections need internet.

Security and data boundaries

05Against the field

Most memory products are an API. This one is a folder.

Mem0 and Zep are strong choices for developers building memory into their own product. Built-in vendor memory is fine if you only ever use one assistant. Central Brain is for the person who uses several tools and wants one memory they own.

Criterion	Central Brain	Mem0	Zep	ChatGPT memory
Where memory lives	Storage you choose. Plain files plus a local index.	Mem0's hosted store (self-host with the open source version).	Zep cloud, or your own cloud on Enterprise.	Your OpenAI account.
Can you read it yourself	YESOpen the folder.	VIA APIAI-extracted summaries.	VIA APIGraph nodes and facts.	PARTLYAuto-rewritten summary since June 2026.
Works across tools	Claude Desktop, Claude Code, GitHub Copilot and the ChatGPT desktop app, over MCP.	Apps you build with the SDK. OpenMemory MCP for coding tools.	Apps you build with the SDK. Memory MCP Server seats on every plan, one on Free.	ChatGPT only.
Checkpoints and recall	YESmemory_checkpoint, then memory_recall in the next session.	NOAutomatic extraction instead.	NOAutomatic graph updates instead.	NO
Pricing model	Flat per license. $12 a month for one.	Free tier, then $19 and $249 a month with retrieval quotas.	Free credits, then $125 and $375 a month, metered per 350 bytes.	Included in your ChatGPT plan.
Best for	People and small teams who use several AI tools and want one memory they own.	Developers embedding memory in their own product.	Enterprise agent teams that need governance and graph retrieval.	Staying inside ChatGPT.

Competitor details checked September 17, 2026 against mem0.ai/pricing, getzep.com/pricing, TechTimes on the ChatGPT Dreaming update and Notis on ChatGPT memory export.

06Pricing

Flat per license. No tokens, credits or retrieval quotas.

Pay monthly, or once a year for 15% off. Every plan includes the local-first workflow, project memory over MCP, semantic and exact search, and the desktop app for Windows and Apple silicon Mac.

Solo

$12/ month

One Central Brain license. $12 per license.

Local-first memory and files
Project memory over MCP
Semantic and exact search
Windows and Apple silicon Mac

Get Solo

Lowest per license

Power Pack

$35/ month

Four licenses. $8.75 each.

Everything in Solo
Four separate one-computer keys
Best value for a small team

Get Power Pack

Team and Enterprise

Let's talk

Infrastructure and support sized to your team.

Custom pricing
Deployment support
Ask about Pinecone or Postgres pgvector

Talk to NeuroAIgent

Annual plans are billed once per year at 15% below twelve monthly payments. Taxes may apply. See the refund and cancellation policy.

07AEO Presence

Knowledge in. Knowledge found.

Central Brain helps your AI tools remember what you know. AEO Presence makes sure the answer engines remember you exist. It scores any site across technical SEO, on-page SEO, content quality, answer engine readiness and social sharing with 25+ deterministic checks, then ranks the fixes by impact.

Explore AEO Presence Learn AEO

26%of the score is answer engine readiness: structured data, question headings, author signals, llms.txt

25+deterministic checks on live HTML. Same input, same score, every time

1:1head-to-head comparison against a competitor's site

08Straight answers

Frequently asked questions.

More answers on the FAQ page, and every term defined in the glossary.

What is Central Brain?

Central Brain is a local-first AI memory layer by NeuroAIgent: a Windows and Apple silicon Mac desktop app that indexes project files locally and gives supported AI tools, such as Claude Desktop and GitHub Copilot, persistent project memory and source-traceable search over MCP (Model Context Protocol). It prepares PDF, DOCX, TXT, Markdown, CSV and XLSX files and searches them with a bundled embedding model. Claude Desktop, Claude Code, GitHub Copilot and the ChatGPT desktop app connect on Windows and Apple silicon Macs. Preparing files, indexing, search and memory run without a cloud AI provider, and licensing uses Keygen. See the product page.

Is Central Brain a second brain?

Yes, in the practical sense: it is a store of your files and project notes that you and your AI assistants can both read. The MCP server keeps a per-project memory of five note types (decisions, facts, questions, todos and summaries) plus a project brief, saved as local files. Your assistants can read it at the start of a session and add to it as work happens. Unlike a notes app, the store is searchable by meaning and reachable from Claude Desktop, Claude Code, GitHub Copilot and the ChatGPT desktop app. Unlike built-in chat memory, it lives on storage you choose. Read more at second brain for AI.

Who is Central Brain for?

Central Brain is for people who work with an AI assistant across many sessions on the same projects and are tired of re-explaining context. That includes developers using Claude Code, GitHub Copilot or Codex in the ChatGPT desktop app on a codebase, and consultants and researchers with folders of PDFs and spreadsheets. It also includes small teams who want one memory per project rather than one per person per tool. It suits anyone who needs documents kept local. It is a poor fit if you need a hosted memory API to call from your own application; for that, look at Mem0 or Zep.

Does Central Brain replace ChatGPT or Claude?

No. Central Brain does not include a chat model and does not answer questions itself. It sits beside the assistants you already use and gives them memory and file search. Claude Desktop, Claude Code, GitHub Copilot and the ChatGPT desktop app connect to it over MCP (Model Context Protocol), call its tools, and do the reasoning and writing. You keep your existing subscriptions. Central Brain adds the part those tools lack: a local, inspectable, shared memory and a source-traceable index of your files.

How is Central Brain different from built-in ChatGPT or Claude memory?

Three differences. Location: built-in memory lives in one vendor's account; Central Brain memory lives in files on your machine. Portability: ChatGPT memory works only inside ChatGPT and is not included in data export (source). Claude Code memory is scoped to a project through CLAUDE.md. Central Brain memory is read from one store by Claude Desktop, Claude Code, GitHub Copilot and the ChatGPT desktop app. Control: ChatGPT memory is capped and, since June 2026, auto-rewritten into a summary you cannot edit line by line. Central Brain notes are plain files you can open. Learn how these approaches differ in the AI memory layer guide.

What does "central memory" mean?

Central memory means one memory store per project that every connected AI tool reads from and writes to, instead of separate memories inside each app. When you make a decision in a Claude session, a checkpoint goes to the project's memory files. When you open Copilot or Codex on the same project, memory_recall can return that decision. The person and their agents share the same notes. The store is on your disk, so it is also central in the sense of being under your control rather than spread across vendor accounts.

The AI doesn't own the memory. You do.

Central Brain runs on Windows and macOS (Apple silicon). In the default workflow, files, prepared output, vectors and memory stay on storage you choose.

Get Central Brain Read the setup guide

AlsoFrom NeuroAIgent

Your AI forgets everything. Central Brain™ doesn't.

Every session starts from zero. Or worse, from a summary a model wrote about you.

Compaction erases the middle of your work.

Vendor memory is a retention feature.

The same files, uploaded again and again.

Tools that shut down take your memory with them.

Five kinds of notes and a project brief. One place for every connected tool.

Five steps. Then the memory compounds.

Choose

Prepare

Index

Connect

Refresh