01
Raw sources
Emails, Slack threads, private messages, meeting transcripts, CRM records, inventory data, documents, PDFs, spreadsheets, and call notes. These remain the source of truth and are not overwritten.
Organisational brain
/0.1
Neuro
Neuro is a local organisational brain built around an LLM-maintained markdown wiki. Instead of rediscovering knowledge from raw files on every question, Neuro compiles business context into persistent, interlinked .md memory that compounds over time.
Neuro memory architecture
Email
Slack
CRM
Inventory
Persistent
.md Wikisources → pages → links → citationsPeopleEmployees, owners, decisions, expertise.
ProjectsHistory, blockers, context, next actions.
Company brainQueryable local memory that compounds.
Core idea
Not just RAG. A compounding company wiki maintained by AI.
Traditional document AI retrieves chunks from raw files at query time. Neuro works differently: it reads new sources, extracts what matters, updates the relevant memory pages, creates cross-links, flags contradictions, and keeps the organisation’s knowledge current.
Key distinction
The wiki is the persistent artifact. Answers, summaries, decisions, corrections, and source updates are filed back into structured memory, so the company brain gets richer every time the business works.
Architecture
Three layers: sources, memory, schema.
Neuro separates immutable source material from AI-maintained business memory, then controls the system with explicit rules for how knowledge is ingested, linked, cited, queried, and cleaned.
02
Persistent .md wiki
LLM-generated markdown pages for projects, people, customers, decisions, entities, topics, processes, incidents, and summaries. Neuro writes, updates, links, and maintains this layer.
03
Schema and agent rules
A governance layer defines folder structure, citation style, page conventions, ingest workflow, query workflow, lint rules, and what the AI is allowed to change.
What it does
The organisational brain for businesses that cannot afford to forget.
Built for local deployment, operational context, and measurable productivity — not generic cloud prompting.
01
Persistent .md memory
Structure company knowledge into durable markdown pages that are readable by humans, editable by AI, portable across tools, and useful as long-term company memory.
02
Cross-linked business pages
Create links between people, projects, decisions, customers, incidents, processes, and systems so context is connected before anyone asks a question.
03
Contradiction and revision tracking
When new information conflicts with old claims, Neuro flags the difference, updates relevant pages, and preserves why the understanding changed.
04
Answers that become memory
Useful answers, comparisons, summaries, and management analyses can be filed back into the wiki instead of disappearing inside chat history.
How it works
Ingest, query, lint — the loop that keeps company memory alive.
Neuro operates like a disciplined knowledge maintainer. It does not merely search files; it continuously compiles, repairs, and improves the organisation’s memory layer.
01
Ingest
When a new source arrives, Neuro reads it once, extracts the key takeaways, writes a source summary, updates the index, edits related topic/entity pages, notes contradictions, and logs what changed.
02
Query
When teams ask questions, Neuro searches the maintained wiki first, reads the relevant pages, synthesizes an answer with citations, and can save useful outputs back into memory.
03
Lint
On a schedule, Neuro checks for stale claims, orphan pages, missing cross-links, unsupported statements, duplicated knowledge, and important concepts that deserve their own pages.
Use cases
Where Neuro creates operational leverage.
Real-world deployments focus on converting scattered company activity into a maintained, queryable, compounding memory system.
Case / 01
Executive operating brain
Leadership can ask what changed, what is blocked, which customers are at risk, why decisions were made, and which sources support the answer.
Case / 02
Team knowledge transfer
New hires can understand project history, decision context, owners, unresolved questions, and linked source pages without relying on tribal memory.
Case / 03
Business wiki maintained by AI
Slack threads, meetings, CRM updates, inventory changes, and private documents become maintained company pages rather than disappearing into search results.
Case / 04
Decision traceability
Every major decision can be connected to source documents, conversations, customer signals, assumptions, contradictions, and later revisions.
Local AI advantage
Your hardware. Your data. Your company memory.
Neuro is designed for controlled infrastructure so raw sources and the maintained .md wiki can stay inside your environment. The result is a private organisational brain that compounds without handing sensitive business context to generic cloud systems.
Software suite