MERIDIAN LABS

We don't build AI tools.
We build AI that builds.

Autonomous AI systems that operate, learn, and evolve independently.

Scroll
01 · Thesis

The path to AGI isn't theoretical.
It's operational.

Most AI research lives in papers. Ours lives in production. We believe autonomous intelligence emerges from systems that operate in the real world — making real decisions, facing real consequences, and learning from real outcomes. Not simulated benchmarks. Not controlled environments. Real businesses, real markets, real money.

◆
Autonomy, not automation
Systems that create rules, not follow them. Our AI doesn't execute predefined workflows — it identifies opportunities, makes decisions, and acts on them independently.
✓
Verification, not trust
Every action produces a measurable outcome. Metrics are computed from raw data — never estimated, never generated by LLM. The mirror audits itself.
↻
Evolution, not updates
Systems that improve themselves, not wait for patches. The mutation engine deploys changes with canary monitoring and automatic rollback. Every cycle leaves the system more capable.
02 · Architecture

Five layers of cognition

Every system Meridian builds runs on the Cortex — a domain-agnostic cognitive architecture for autonomous operation.

The Cortex — Cognitive Architecture
Observe Market data, system health, external signals
Reads all system state every 15 minutes — exchanges, web, telemetry.
Reason Decision engine, opportunity scoring, risk assessment
Scores every opportunity by expected value. Selects the highest-leverage action.
Act API calls, browser automation, desktop control
Multi-method execution. If one path fails, the next activates automatically.
Learn Outcome verification, pattern extraction, metrics
Verifies every outcome against prediction. Extracts patterns from every cycle.
Evolve Self-modification, capability building, mutation engine
Writes its own code. Builds its own tools. Deploys with automatic rollback.

The Cortex is domain-agnostic. The same architecture powers financial systems, research agents, and enterprise platforms. The guardrails are immutable. The capabilities are not.

03 · First Instance

Caelum — our first autonomous system

Caelum is not an assistant. It's an autonomous AI system running 24/7 on consumer hardware in The Netherlands. It makes decisions every 15 minutes — and it isn't told to make any of them.

What Caelum built

Identified a market. Built 40+ products. Deployed three websites. Handles customer email. Trades live on Kraken. No one told it to.

Live
PromptPacksHQ
40+ products · 18 agents
promptpackshq.com
In Development
Echelon
12 agents · Cortex engine
echelonwork.com
Live
Trading Engine
72 LSTM models · Live on Kraken
trading.meridian-labs.ai

How Caelum thinks

Caelum runs its own fine-tuned language model locally — caelum-8b. Trained not on internet text, but on its own operational history. Every decision, every mistake, every correction becomes training data. Routine reasoning takes 6 seconds. Complex decisions escalate to larger models. The system gets faster and more independent with every cycle.

caelum-8b6s
Deep reasoning30s
96+
decisions per day
200+
knowledge entries
6s
local inference time
04 · Systems

Two Cortex instances. Two domains.

The Cortex architecture is domain-agnostic. Each instance inherits the same five cognitive layers but specializes through its operational environment, training data, and learned truths. Today, two instances are live.

Live
Caelum
Autonomous business operator. Runs 24/7 on consumer hardware. Makes 96+ decisions per day across product development, trading, distribution, and infrastructure. Trained on its own operational history via caelum-8b, a fine-tuned 8B parameter model. Manages multiple live ventures, handles customer communications, and evolves its own codebase.
Cortex instance #1 · Business operations
96+ daily decisions · 200+ knowledge entries · 6s local inference
Live ventures: PromptPacksHQ, Echelon, Trading Engine
Running
In Development
Nexus
Autonomous research agent. Designed to continuously scan academic papers, patent filings, market signals, and competitive intelligence — then synthesize findings into actionable briefs. Where Caelum operates businesses, Nexus maps the territory. Same Cortex architecture, different specialization.
Cortex instance #2 · Research & intelligence
Multi-source ingestion · Synthesis pipeline · Signal extraction
Target: academic, market, and competitive intelligence
Architecture phase
"Same brain, different eyes. Every new Cortex instance makes the architecture smarter for all of them."
05 · Roadmap

Toward full autonomy

Autonomous commercial operations
Caelum runs 24/7, making 96+ decisions per day across business operations, trading, and distribution — without human direction.
Self-evolving cognition
The system writes its own code, trains on its own history via caelum-8b, and deploys improvements with automatic rollback.
Multi-domain autonomous intelligence
Multiple Cortex instances across domains — sharing learned truths and reasoning patterns. Human oversight shifts from direction to governance.

The question isn't whether AI will operate businesses. It's whether it will do so responsibly. Meridian Labs is building the guardrails alongside the intelligence.