🧠 Intelligence Layer · Memoria

Your AI team's long-term memory — structured, alive, and always connected.

Memoria goes beyond a document store. It's a living knowledge graph where every vault note, KB document, and agent insight lives as a node — connected by semantic edges, queryable by your entire AI team, and intelligently maintained through confidence scoring and decay.

Knowledge Base vs Memoria — Two different layers, two different jobs. The Knowledge Base holds documents that agents search for RAG (PDFs, DOCX, Drive files). Memoria holds structured knowledge — facts, decisions, people, risks, agent insights — as graph nodes with typed connections between them. Agents use both, but they serve fundamentally different purposes.

Knowledge Base vs Memoria — at a glance

Capability Knowledge Base Memoria
What it stores Documents — PDF, DOCX, Drive Structured nodes — facts, insights, entities
How agents use it Semantic search (RAG) Graph query + semantic edges
Auto-populated ❌ Requires upload ✅ Crystallisation from every conversation
Confidence over time ✅ Decays & reinforces
Obsidian / vault import ✅ ZIP/Markdown with YAML frontmatter
Graph visualisation ✅ Force-directed D3 graph
Health monitoring ✅ Orphans, duplicates, stale, low-confidence

How Memoria works — step by step

1
Understand the two knowledge layers

Before ingesting anything, understand how AEGIS OS organises knowledge. The Knowledge Base is a flat document store — you upload files, they get chunked and indexed for semantic search. Memoria is a graph — nodes (facts, entities, decisions, insights) are connected by typed edges (related_to, crystallized_from, contradicts, references). The same fact can live in both layers and serve different retrieval paths.

The two knowledge layers
📚 Knowledge Base
PDFs · DOCX · Drive
16 documents
142 chunks indexed
→ RAG retrieval
🧠 Memoria
Vault notes · Agent insights
5,203 nodes · 599 edges
→ Graph query
Both layers are used by all 8 agents simultaneously.
2
Upload a vault or single Markdown file

Go to Memoria Explorer (sidebar → 🧠 Memoria). Drag a ZIP of your Obsidian vault, or drop a single .md file. AEGIS reads YAML frontmatter — tags, aliases, links — without calling any LLM. Your pre-tagged notes are ingested in seconds. For untagged notes, AEGIS runs LLM enrichment as a background job, batching up to 5 notes per second.

Tip: If you use Obsidian with the Dataview or Templater plugins, run your local pipeline first to pre-populate tags: in every note's frontmatter. The ingest will be instant and zero LLM calls needed.

Vault upload — drag & drop
🗂️
my-obsidian-vault.zip
5,196 Markdown files · ZIP archive
✅ 5,137 nodes tagged from YAML frontmatter — 3.2s
5,137
Tagged (YAML)
59
Queued for LLM
599
Edges created
3
Explore the knowledge graph

Once ingested, the Graph Explorer renders a live force-directed D3.js graph. Nodes are colour-coded by type: blue = vault notes, green = KB documents, orange = agent insights. Node opacity reflects confidence — fully opaque = high confidence, faded = decaying. Superseded nodes are outlined in red.

Click any node to open the detail panel — see its full content, tags, confidence score bar, linked nodes, and an option to approve, reinforce, or supersede it. Use the type filter chips and search bar to focus the graph on what you need.

Knowledge Graph Explorer
core insight KB doc KB doc superseded
Vault notes (opacity = confidence)
KB documents
Agent insights
Superseded
4
Let agents crystallise insights automatically

Every time an agent responds to a conversation, Crystallisation runs in the background. It extracts 2–3 key insights from the response, saves them as orange agent_insight nodes, and creates semantic edges linking them to related existing nodes. No manual work required. The graph grows smarter with every conversation.

Boardroom sessions are always crystallised — even if you've hit the per-tenant rate limit for regular chats. High-quality crystallisation requires responses of at least 350 characters and excludes short acknowledgements ("OK, got it", "Sure").

Recently crystallised — agent insights
🔮 Angie · Chief of Staff · Crystallised 12 min ago · 94% confidence
TechCorp Group require a signed GDPR Data Processing Agreement before approving any final proposal. Legal review is a hard prerequisite for close.
🔮 Viktor · Finance · Crystallised 2h ago · 81% confidence
Q1 runway is 14 months at current burn rate. No external financing needed before Q4 unless major deals stall for 60+ consecutive days.
🔮 Aria · Marketing · Crystallised 8 days ago · 41% confidence ⚠️
LinkedIn Ads CPL dropped 18% after switching to Thought Leadership format in week 3. Recommend scaling budget.
Faded node = confidence decaying below 40% — confirm or supersede.
5
Understand confidence scoring and decay

Every agent_insight node starts at 0.6 confidence. Human-authored nodes start at 1.0 and never decay. Agent insights decay approximately 5% per month — a fact not confirmed in 6 months drops to ~0.75, and in 18 months to ~0.6. When you or another source confirms the same fact, the node is reinforced (+0.15, capped at 1.0) and the clock resets.

If a fact is replaced by a newer version, you can supersede it — the old node's confidence drops to 0.1 and is linked to the new one. This preserves history while making clear which version is current.

Confidence scores — node detail panel
TechCorp GDPR DPA prerequisite
94%
Q1 runway 14 months at current burn
81%
LinkedIn Ads CPL drop 18% in wk3
61%
Competitor X launched free tier in Feb
34%
Green ≥ 70% · Yellow ≥ 40% · Red < 40% · Decay: ~5%/month on agent insights
6
Run a Health Check — keep the graph clean

Click Run Health Check in the Memoria Explorer sidebar. AEGIS scans the full graph for six classes of issues and returns a health score from 0–100. Aim for 80 or above. A fresh vault import will typically score 70–80; after a week of crystallisation and a few enrichment passes it will reach 85+.

87
Health Score
5,203 nodes · checked just now
Orphan nodes ✓ Clean
Untagged notes ✓ Clean
Isolated clusters 2 found
Potential duplicates 3 found
Stale nodes ✓ Clean
Low confidence nodes 12 found

When to use Memoria

🗃️
Obsidian vault import
Bring your personal knowledge base into AEGIS — your AI agents now know what you know. Tags, links, and structure preserved.
🧠
Long-term agent memory
Agent insights compound across conversations — facts confirmed multiple times gain confidence, contradictory info is flagged.
🔍
Cross-domain context
When a sales insight connects to a finance risk which links to an ops bottleneck — Memoria surfaces the chain automatically.
📊
Decision audit trail
Every Boardroom synthesis crystallises into Memoria. Build a searchable history of every major decision and why it was made.
⚠️
Stale knowledge detection
Competitor intel, pricing assumptions, and market facts decay over time. Health Check flags anything that needs re-verification.
🔗
Semantic linking
Automatically connect related nodes across different domains — a client note linked to a contract node linked to a risk insight.

Frequently asked questions

What's the difference between the Knowledge Base and Memoria?
The Knowledge Base holds documents (PDFs, DOCX, Google Drive files) that are chunked and indexed for semantic search — agents retrieve relevant passages via RAG. Memoria is a structured graph of nodes — facts, decisions, entities, and agent insights — connected by typed semantic edges. Agents use both: KB for document retrieval, Memoria for structured context and long-term memory.
Can I import my existing Obsidian vault?
Yes. Zip your vault folder and upload it in Memoria Explorer. AEGIS reads tags:, aliases:, and wikilinks from YAML frontmatter without calling any LLM — your pre-processed notes ingest in seconds. For notes without frontmatter, a background enrichment job will tag them using your LLM configuration.
What is Crystallisation?
Crystallisation is the automatic process that extracts 2–3 key insights from every agent conversation and saves them as graph nodes. It runs after every response in the background — rate-limited to 5 per 10 minutes per tenant to avoid noise. Boardroom sessions always crystallise regardless of the rate limit. Responses under 350 characters or containing only acknowledgements are skipped.
How does confidence decay work?
Agent insights (trust level: agent_proposed) decay approximately 5% per month using an exponential formula: score × 0.9983^(days_since_reinforced). The floor is 0.1 — a node never reaches zero. Human-authored nodes (uploaded by you) start at 1.0 and do not decay. When a fact is confirmed by another source, reinforce the node to add 0.15 (capped at 1.0) and reset the decay clock.
What does the Health Check measure?
Six checks: Orphan nodes (no connections), Untagged notes (enrichment didn't run), Isolated clusters (groups disconnected from the main graph), Potential duplicates (60%+ title word overlap), Stale nodes (untouched 30+ days), and Low confidence nodes (score below 40%). The score is 0–100, weighted by how much each issue affects graph quality.
Can agents query Memoria directly in a conversation?
Agents automatically receive relevant Memoria context when answering questions — the graph is queried in the background and injected into the agent's context window. You don't need to prompt them specifically. For deep graph exploration, use the Memoria Explorer in the app.
Free to start
Build your team's long-term memory today.

Upload a vault, start talking to your agents,
and watch Memoria build itself — one conversation at a time.

Start free — no card needed →