v0.11 · Local Genesis — Active

Alive.
Approachable.
Dynamic.

The first intelligence-native platform built on LaaP

Cortex One is not a chatbot wrapper. It is a full mesh of semantic nodes — identity, orchestration, knowledge graph, policy engine — working together to make meaning discoverable across your organisation.

Enter Cortex One Run it locally The protocol behind it →
9+
Nodes
Datasources
100%
Data Sovereign
0
Vendor lock-in
The Mesh

Every node has one job. Together they are intelligent.

Cortex One runs as a Docker mesh. Each node is independent, sovereign, replaceable. The intelligence is in the connections — not in any single model.

🪪
Identity Node
LaaP JWT issuance, genesis flow, OIDC bridge. Every participant starts here.
:8400
Orchestrator
Routes natural language intent to the right semantic node. No API contract needed.
:8200
Semantic Node
NL→SQL, capability cards, LLM-agnostic. Owns its data, enforces its policies.
:8100
🔗
Knowledge Graph
Entity reconciliation, relationship mapping, semantic federation across nodes.
:8300
🛡
OPA Policy Engine
Every query evaluated before answering. Zero trust enforced at the edge.
:8181
🔑
Vault Node
Fernet AES-256 encryption for secrets and API keys. Your keys stay local.
:8090
🧠
Cortex Brain
LangChain agent runner with tool registry. Composable AI workflows.
:8700
📡
Cortex Pulse
Eclipse Mosquitto MQTT — the event bus. Typed, signed, real-time.
:1883
🏗
Node Builder
Create new semantic nodes on-demand. Your data, your domain, your rules.
:8500
What makes it different

Built on principles, not just patterns.

Cortex One shows what becomes possible when you build with LaaP as the foundation from day one.

"Cortex One shows what becomes possible with LaaP. It is one implementation — not the only one."
LaaP Communication Guidelines
🔒
Data never moves. Only meaning does.
Each semantic node owns its datasource. Queries are resolved locally. Only the answer — not the raw data — leaves the node. Full sovereignty by design.
🧠
LLM-agnostic by design.
Run with OpenAI, Ollama, Anthropic or a mock LLM for local development. No vendor lock-in at the inference layer either.
🗣
Natural language as the interface.
Ask your data a question in plain language. The orchestrator routes the intent, the semantic node resolves it, the knowledge graph enriches it.
🪪
Identity is first-class.
Every action is signed with a LaaP JWT. Who asked? What did they ask? What was permitted? Immutably recorded at the node level.
🌐
Open, extensible, yours.
Add a new semantic node for any datasource or capability. The mesh grows with your organisation. No ceiling, no platform tax.
🐳
Runs anywhere Docker runs.
One compose file. Local machine, on-premise server, private cloud. No managed service required. You own the stack.
Local Genesis

Run the full mesh on your machine.

No cloud. No subscription. No data outside your machine. One command to boot the entire intelligence stack locally.

Local Genesis · v0.11

From zero to YOU Born in minutes.

Docker, a terminal, and a few gigabytes of RAM. That is all you need to run a full Cortex One mesh — complete with semantic nodes, knowledge graph, identity node and policy engine.

Visitor Docker Mesh Boot Genesis Owner LaaP JWT YOU Born 🎉
Enter Cortex One Create your identity
The protocol behind it

Cortex One is built on LaaP.

LaaP is the open protocol. Cortex One is one implementation. Understand the foundation — build your own implementation, or extend ours.

Explore the Protocol → Create your identity Read the Manifesto