Central Brain by NeuroAIgent

Central Brain AI: the local-first AI memory layer.

Choose a plan Explore the features ↓ Get Central Brain on the Microsoft Store

Windows + Apple silicon MacLocal MiniLM embeddingsSource-traceable results

What is Central Brain AI?

Central Brain AI is NeuroAIgent’s local-first AI memory layer, a desktop app for Windows and Apple silicon Macs. It indexes project files on your computer, keeps project memory as readable local files, and runs air-gapped between 14-day license checks (cloud-hosted AI clients still need internet). Through a local MCP server with eight tools, Claude Desktop, Claude Code, GitHub Copilot and the ChatGPT desktop app can recall that memory and search your files. The name is sometimes written as one word, CentralBrain; it is the same NeuroAIgent product. Full answer: what is Central Brain AI?

**Keep knowledge under your control.**Local-first processing and storage you choose.

**Stop reprocessing unchanged files.**Refresh the material that actually changed.

**Give AI answers it can trace.**Keep source, page, and chunk context attached where available.

Stop making your AI start over.

The most valuable knowledge in your business is trapped in folders: contracts, reports, spreadsheets, notes, PDFs, and media. Uploading files one conversation at a time is slow, inconsistent, and hard to audit. Central Brain prepares that material once, keeps it organized by project, and makes it available to supported AI tools as a reusable memory layer.

Without Central Brain

• Scattered folders and stale indexes
• The same files uploaded again and again
• Useful answers with no clear source trail
• Every AI conversation starts from zero

With Central Brain

• Prepared, AI-ready project files
• New and changed-file workflows
• Source paths, pages, and chunks preserved where available
• Searchable knowledge your team can reuse

Looking for a private AI memory layer? Start here.

Central Brain is designed for people who need a local-first way to prepare, search, and reuse project knowledge across supported AI tools.

How do I search private PDFs and documents with AI?

Prepare PDF, DOCX, TXT, Markdown, CSV, and XLSX content once, then search the local index without re-uploading the same files for every conversation.

Which local AI memory layer keeps results tied to sources?

Central Brain results include the source path, prepared-file path, page, and chunk where available, so retrieved context keeps a usable trail back to the original material.

Can Claude, Copilot, ChatGPT, or Codex use my project files?

Central Brain provides project-specific MCP setup for Claude Desktop, Claude Code, GitHub Copilot and the ChatGPT desktop app (Chat, Work and Codex modes). Capabilities vary by client.

Can I start locally and move to team infrastructure later?

Yes. Start with bundled local embeddings and JSONL vectors. Optional connections can later publish to Pinecone or Postgres pgvector or JSONB; talk to NeuroAIgent about Team and Enterprise deployments and implementation support.

The Central Brain MCP server: eight tools

Eight MCP tools. File tools are read-only; memory_checkpoint appends notes; memory_compact replaces the project brief and marks older notes as compacted rather than deleting them. On connect, central_brain_ping returns a RESUME block from the project’s saved memory, so a new session picks up where the last one left off. Supported AI tools connect to this local server over the Model Context Protocol (MCP); capabilities vary by client.

central_brain_ping

Confirms the server is running, reports the project root folder and returns a RESUME block from the project’s saved memory.

memory_recall

Searches decisions, facts, open questions, todos, summaries and the project brief. Call it before answering anything about earlier work.

memory_checkpoint

Appends one note at a time (a decision, fact, question, todo or summary), so it is still there after the chat is compacted or closed.

memory_compact

Replaces the project brief with an updated summary and marks older notes as compacted. It does not delete them.

semantic_search

Vector search by meaning over the project's prepared files, using the bundled local MiniLM model by default. Hits carry path, page and chunk where available.

search_text

Exact text search across the files for when you know the phrase.

read_file

Reads one file under the project root.

list_files

Lists files under the root, optionally in a subfolder.

Central Brain MCP server setup guide

Built for the work behind the prompt.

Central Brain handles the unglamorous, essential work that makes project knowledge useful to AI—preparation, enrichment, indexing, retrieval, traceability, and controlled connection.

01 / PREPARE

Make messy files AI-ready

Prepare PDF, DOCX, TXT, Markdown, CSV, and XLSX content, with metadata-only indexing for images, audio, and video. Optional OCR can recover text from scans and photos when it is included in the build.

02 / SEARCH

Search locally by default

A bundled 384-dimensional MiniLM model, local vector storage, and cosine-similarity search keep the core workflow useful without requiring a cloud AI provider.

03 / TRACE

Trace results to the source

Results include the source path, prepared-file path, page, and chunk where available. Useful results do not become mystery snippets.

04 / REFRESH

Update what changed

Review new, changed, embedded, removed, and failed files. Re-embed a selected document or only new and changed material, and resume interrupted work.

05 / CONNECT

Use the AI tools you know

Deploy project-specific MCP access for Claude Desktop, Claude Code, GitHub Copilot and the ChatGPT desktop app. Project memory is kept as five note types plus a project brief, in readable local files.

06 / SCALE

Scale without rebuilding

Start with local JSONL vectors. Optional connections can publish to Pinecone or Postgres pgvector or JSONB when a team needs shared infrastructure; ask NeuroAIgent about Team and Enterprise deployment support.

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

In the default project workflow, your source files, prepared output, vectors, and memory stay on storage you choose, and local MiniLM embeddings are bundled with the app. Central Brain contacts Keygen for activation and a license check every 14 days, and runs air-gapped in between; Keygen receives licensing and device data, not documents. Connected AI tools receive the context they retrieve, and cloud embeddings, Pinecone, Postgres, Redis, or other optional connections send selected data only when you enable them.

• Local credential storage for optional connections
• Project-by-project organization and permissions
• Read-only MCP file tools; memory tools append notes and update the project brief
• Revocable same-Wi-Fi device trust with sharing off by default

Explore Security & Data Boundaries

Five steps. Then your knowledge compounds.

01

Choose

Select source folders and where prepared output should live.

02

Prepare

Convert and enrich files into AI-ready material, with optional OCR where the build includes it.

03

Index

Chunk and embed locally or with a provider you configure.

04

Connect

Search directly or connect supported AI tools through MCP.

05

Refresh

Process only new and changed material as the project evolves.

See how Central Brain works

Give every person the memory they need.

Pay monthly or save 15% with one annual payment. Start local, then add infrastructure integrations and hands-on support when your deployment demands them.

Annual plans are billed once per year.

Solo

One Central Brain license.

$12 / month

$12 per license

• Local-first workflow
• Project search and preparation
• Supported MCP connections Buy now

Pair

Two Central Brain licenses.

$20 / month

$10 each · save $4 monthly

• Everything in Solo
• Two activated licenses
• Built for a working pair Buy now

Power Pack

Four Central Brain licenses.

$35 / month

$8.75 each · save $13 monthly

• Everything in Solo
• Four activated licenses
• Serious value for small teams Buy now

Team & Enterprise

Infrastructure and support sized to your team.

Let’s talk

Custom deployment

• Custom pricing and scope
• Deployment support
• Ask about Pinecone or Postgres Talk to NeuroAIgent

Choose a monthly subscription or an annual subscription billed once per year at 15% below twelve monthly payments. Taxes may apply. Each included license activates one computer. Review the refund and cancellation policy; renewal and cancellation terms are shown at checkout. Team and Enterprise scope is defined in your order.

Frequently asked questions.

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, Claude Code, GitHub Copilot and the ChatGPT desktop app, persistent project memory and source-traceable search over MCP. It prepares, indexes, and searches PDF, DOCX, TXT, Markdown, CSV, and XLSX files, and keeps project memory as readable local files: five note types (decision, fact, question, todo, and summary) plus a project brief.

How can I search private documents with AI without uploading them in every chat?

Central Brain prepares files once, stores the default vector index locally, and lets supported AI tools retrieve relevant project context through a project-specific MCP connection. You control the source folders, prepared output, and optional cloud connections.

Is Central Brain a private RAG tool?

Central Brain provides the document preparation, local embeddings, vector search, source metadata, and MCP access used in private retrieval workflows. The connected AI assistant generates the final answer, while Central Brain supplies relevant context and its source trail.

Can I move my license to a replacement computer?

Yes. Deactivate the old computer first when possible, then activate the same key on the replacement. Each self-service license includes three complimentary assisted machine resets per rolling 12 months when the old device cannot be deactivated. Transfers require an active paid subscription and remain subject to the one-computer limit.

Who is Central Brain for?

Central Brain is designed for developers, consultants, researchers, and small teams who reuse project files across AI sessions and want reusable AI context without making a cloud service the default home for every source file.

Does Central Brain replace ChatGPT, Claude, Copilot, or Codex?

No. Central Brain is the memory layer behind supported AI tools. It prepares and retrieves your project material; the connected assistant remains the interface that reasons over and responds with that context.

What stays local?

In the default workflow, files, prepared output, vectors and memory stay on storage you choose. Keygen receives licensing and device data, not documents. Connected AI tools receive the context they retrieve, and optional cloud connections send selected data only when enabled.

Does Central Brain work offline?

Central Brain runs air-gapped: preparing files, indexing, search and memory need no network connection. The license is re-checked online every 14 days, and Team and Enterprise plans can get a build with no online license check. Optional cloud connections and cloud-hosted AI clients need internet.

What file types does it support?

PDF, DOCX, TXT, Markdown, CSV and XLSX, with metadata-only indexing for images, audio and video; OCR is optional and depends on the build.

Which AI tools can connect?

Central Brain provides project-specific MCP setup for Claude Desktop, Claude Code, GitHub Copilot and the ChatGPT desktop app (Chat, Work and Codex modes) on Windows and Apple silicon Macs. Setup and capabilities vary by client.

Can my team use Pinecone or Postgres?

Pinecone and Postgres (pgvector or JSONB) are optional connections for publishing the index to shared infrastructure, and they send selected data only when enabled. Contact NeuroAIgent to discuss which plan, deployment, and support fit your team.

Does Central Brain guarantee an AI answer?

No. Central Brain prepares and searches source material and exposes relevant project context. Outputs from third-party AI systems remain subject to those systems and should be reviewed.

Your best work should compound—not disappear into another folder.

Build a memory layer your AI can reuse, your team can inspect, and your business can keep under control.