Your companion for governed financial intelligence.
FI through AI, powered by SI.

🇿🇦 Built in South Africa. · POC before narrative · receipts before claims · time reveals

---

👋 Meet LEFA

_{Canonical source → animated interface interpretation. The drawing owns identity; the interface may evolve around it.}

Most finance products introduce themselves with charts, balances and buttons.

LEFA starts with a companion.

You speak to LEFA. LEFA is the user-facing base intelligence. The complicated architecture underneath exists to help LEFA make a better decision — not to make the user learn the architecture.

You bring the human question. LEFA brings governed financial intelligence.

---

👁️ Observe → 📖 Ledger → ✨ Reveal

LEFA does not treat a convincing first answer as reality.

Observe what is happening. Ledger what was known. Let time continue. Reveal what survived.

The frontend tells the story. The backend preserves the truth. Time decides what survives.

---

🧠 What moves under LEFA?

The public experience stays light. The backend is intentionally heavy.

HUMAN
  ↓
LEFA — base identity / final decision-maker
  ↓
CRUD — capture the event
  ↓
ARK — governed structured bloat
  ↓
BMP — stress-test, filter and compress
  ↓
MAO — route bounded responsibility
  ↓
SWFUS — five internal ecosystems
  ↓
EVIDENCE + RECEIPTS + UNCERTAINTY
  ↓
LEFA
  ↓
DECISION / EXPLANATION

The internal system does not replace LEFA's judgment. It filters unsupported certainty, preserves useful ambiguity, and returns better governed evidence to the base model.

OPEN // Why the backend expands before it compresses

Human language can carry emotion, memory, contradiction, testimony, uncertainty and several valid meanings at once.

The ARK gives that ambiguity room to become structured bloat instead of prematurely flattening it into one intent. BMP then earns compression by stress-testing what should survive.

HUMAN AMBIGUITY
      ↓
GOVERNED EXPANSION
      ↓
POC / FOC PRESSURE
      ↓
BMP FILTER + COMPRESSION
      ↓
BOUNDED DATA

POC: what can actually be supported.

FOC: what merely looks complete, plausible or polished.

---

🌱 Five agents. Five ecosystems. One LEFA.

The five internal agents do not talk to the user. They receive governed information from LEFA's backend and operate inside bounded ecosystems.

Lane	Ecosystem	Core concern
S	Sovereign Ingestion	What may enter as governed signal?
W	Witness Isolation	What must be preserved independently as testimony?
F	Fluid Vectoring	What interpretations or directions remain plausible?
U	Unified Synchronization	What accepted state must align before action?
S	Severance Execution	What must be cut, held or rejected?

Rich at the center. Minimal at the edges.

The agent receives its identity, role, hierarchy and boundary. It does not need the entire city. It needs its ecosystem.

---

🧊 POC vs FOC

LEFA grows through one loop:

POC → TEST → FEEDBACK → IMPROVE → NEXT POC

A model can sound intelligent and still be wrong. A market can disagree. A beautiful interface can look finished while the backend remains unproven.

So LEFA keeps receipts.

OPEN // The hackathon proof

CONNECT TO ALPACA PAPER ENVIRONMENT
                 ↓
OBSERVE REAL ACCOUNT / MARKET CONTEXT
                 ↓
PRESERVE SOURCE + TIME + PROVENANCE
                 ↓
LEFA RECEIVES GOVERNED FINDINGS
                 ↓
MAKE / EXPLAIN ONE GOVERNED DECISION
                 ↓
WAIT OR REPLAY TIME
                 ↓
COMPARE THESIS WITH OUTCOME
                 ↓
REVEAL

We are not claiming autonomous finance in one sprint.

The POC is smaller and harder to fake:

Can LEFA observe reality, preserve what it knew, make a bounded judgment, and remain accountable to what happens next?

---

🎨 Heavy architecture. Light interface.

LEFA's visual identity is not decoration attached to a dashboard. The companion is the interface anchor.

• black, white and gold;
• calm, recognizable identity;
• circular / halo framing;
• motion maps to system state;
• financial truth must come from real providers, never invented UI values;
• the interface should become simpler as the backend becomes stronger.

Asset governance lives in ./assets/INDEX.md.

---

🛠️ Current project seed

lefa-ai/
├── assets/
│   ├── companion/
│   │   ├── lefa-companion-root.jpg  # canonical drawing source
│   │   └── lefa-companion-root.svg  # animated interface interpretation
│   ├── readme/
│   │   ├── meet-lefa-readme-hero.svg
│   │   ├── lefa-control-room.svg
│   │   ├── lefa-observe-ledger-reveal.svg
│   │   └── lefa-swfus-ecosystem.svg
│   └── INDEX.md
├── docs/
├── src/lefa/
├── tests/
├── .github/
├── .env.example
├── pyproject.toml
├── LEFA AI Logo.png
└── README.md

The code remains deliberately small while the architecture is being validated.

---

⚡ Run LEFA

python -m venv .venv
source .venv/bin/activate
pip install -e '.[dev]'
cp .env.example .env
pytest

CLI:

lefa

Keep credentials in your local .env. Never commit them.

---

🏁 The question

LEFA is not being built merely to answer:

"Can an LLM place a trade?"

The question is:

Can one human-facing intelligence receive messy human intent, use governed internal agents and real financial evidence, preserve uncertainty instead of inventing certainty, and make a decision that time can later validate?

OBSERVE → LEDGER → REVEAL
FI through AI, powered by SI.
Heavy architecture. Light interface.