Sovereign AI agent with 4-store local memory, Aeon DNA identity, and SHA-256 Civic Ledger — LLM-agnostic, zero Docker, runs on your hardware
Project description
vecanova-zana
Your AI doesn't remember your notes. Not a bug. Their business model.
ZANA is a sovereign cognitive agent that lives on your machine, remembers everything, and belongs to you — forever.
The problem
Every AI session starts from zero. You re-explain your context, your preferences, your goals — every single time. Your "personal AI" remembers nothing between sessions. That's not a limitation. That's the design. They want your data on their servers.
ZANA inverts this with one architectural principle: the memory stays with you.
Quick start
pip install vecanova-zana
zana init
zana chat
That's it. No Docker. No cloud account. No API key required to start.
What makes ZANA different
4-Store Memory Architecture
ZANA maintains four types of memory simultaneously — all stored locally in ~/.zana/:
| Store | What it holds |
|---|---|
| Semantic | Concepts, knowledge, domain expertise |
| Episodic | Conversations, events, your history |
| Procedural | How you solve problems, your workflows |
| World Model | Your mental map of projects, people, relationships |
Aeon — Your Sovereign Agent
Every ZANA installation generates an Aeon: a computational identity with 35 genes derived from your usage patterns and a nanosecond-seeded entropy signature. No two Aeons are identical — not even if you reinstall.
Your Aeon evolves through 9 stages (Huevo → Fresh → Rookie → Champion → Ultimate → Legend → Archon → Mythic → Sovereign) as it absorbs your knowledge.
Civic Ledger — Verifiable Sovereignty
Every reasoning decision ZANA makes is signed with SHA-256 and written to ~/.zana/civic_ledger.jsonl. Open it. Read it. Verify it. Your agent can't lie to you about what it decided.
cat ~/.zana/civic_ledger.jsonl | python3 -m json.tool | head -40
LLM-Agnostic by Design
Works with Claude, GPT-4, Gemini, Llama, Mistral, and any OpenAI-compatible endpoint. Switch models with one env var — your memory doesn't move.
export ZANA_PRIMARY_MODEL=claude-opus-4-7 # or gpt-4o, gemini-pro, llama3, ...
zana chat
Zero-Docker, Local-First
Runs on your machine as a standard Python process. SQLite for episodic memory. No containers, no cloud sync, no telemetry.
The Coliseo — Where Knowledge Becomes Combat
ZANA includes a battle engine that maps your Obsidian vault into a 6-world arena:
| World | Domain |
|---|---|
| 🔥 Ignis | Creativity, passion, vision |
| 🔢 Calculus | Logic, mathematics, systems |
| ⚒ Forge | Engineering, architecture, craft |
| 🌊 Oceanus | Strategy, flow, adaptation |
| 🌿 Sylvan | Growth, learning, philosophy |
| ∅ Void | Abstraction, chaos, emergence |
Every note you write in Obsidian shapes the terrain where your Aeon competes. Your knowledge base becomes a living arena.
Z-Skill Protocol
Skills are SKILL.md files that extend ZANA's capabilities. Drop one in ~/.zana/skills/ and ZANA activates it automatically.
zana skill list # see active skills
zana skill install # install from registry
Architecture
~/.zana/
├── aeon_dna.json # Your agent's genetic identity
├── episodic.db # SQLite episodic memory
├── semantic/ # Vector knowledge store
├── procedural/ # Learned workflows
├── world_model.json # Your mental map
├── civic_ledger.jsonl # SHA-256 audit trail
└── skills/ # Active Z-Skills
The core is Python. The security layer (Armor) is Rust — compiled via PyO3, zero overhead.
Configuration
# ~/.zana/.env
ZANA_PRIMARY_MODEL=claude-opus-4-7
ZANA_VAULT_PATH=/home/you/Obsidian/MyVault
ZANA_LANGUAGE=en
Full configuration reference: github.com/Kemquiros/zana-core#configuration
Requirements
- Python 3.12+
- Linux, macOS, or Windows (WSL recommended)
- 512 MB RAM minimum (2 GB recommended for large vaults)
Links
- Landing: zana.vecanova.com
- GitHub: github.com/Kemquiros/zana-core
- Issues: github.com/Kemquiros/zana-core/issues
Your data. Your hardware. Your soul.
JUNTOS HACEMOS TEMBLAR LOS CIELOS.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file vecanova_zana-3.0.1.tar.gz.
File metadata
- Download URL: vecanova_zana-3.0.1.tar.gz
- Upload date:
- Size: 134.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.12.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
b46d2f11da2899881ea499b65b4eb829c5b7d25eb579ec1e0e19d649dd8dabbf
|
|
| MD5 |
e4cba6e85c861196939501d947ae3c8a
|
|
| BLAKE2b-256 |
29c0a2f879a8f406d2a966f31ab1110664a7dd9159d5c7604f3b121563d283aa
|
File details
Details for the file vecanova_zana-3.0.1-py3-none-any.whl.
File metadata
- Download URL: vecanova_zana-3.0.1-py3-none-any.whl
- Upload date:
- Size: 122.8 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.12.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
1b4758bf6d417396513a66cd6ffffecea2c5cb7d0880888348480e8f0d2a2faf
|
|
| MD5 |
8d105edf7ddb054b982cba1f3bb77cb9
|
|
| BLAKE2b-256 |
a9a44e2bc341b22cb0b2bb63ee35decd495fdc330079f53dba902cff250a5d35
|