Skip to content

Personal AI exocortex — multi-agent orchestration + coordination protocol. Monorepo: Core, Forge (orchestration), Quorum (coordination). 9,300+ tests.

License

Notifications You must be signed in to change notification settings

AreteDriver/animus

Repository files navigation

Animus

CI CodeQL Security PyPI - convergentAI License

Personal AI exocortex with production-grade multi-agent orchestration.

Four independently installable packages: a personal AI interface, a workflow orchestration engine, a decentralized agent coordination protocol, and a system bootstrap daemon. 9,600+ tests. v2.0.0.


Quickstart

git clone https://github.com/AreteDriver/animus && cd animus
pip install -e packages/core -e packages/forge -e packages/quorum
pytest packages/core/tests/ packages/quorum/tests/ -q  # 2,600+ tests in seconds

To run the Forge test suite (must run from its package directory):

cd packages/forge && pytest tests/ -q  # 6,700+ tests

What This Does

Animus breaks complex tasks into multi-agent workflows with built-in cost controls, quality gates, and checkpoint/resume. Agents coordinate through a shared intent graph instead of expensive supervisor patterns — O(n) reads instead of O(n^2) messages.


Architecture

┌──────────────────────────────────────────┐
│              ANIMUS CORE                  │
│  Identity · Memory · Learning · CLI       │
│  import animus          1,736 tests  95%  │
├──────────────────────────────────────────┤
│              ANIMUS FORGE                 │
│  Workflows · Budgets · Quality Gates      │
│  import animus_forge    6,731 tests  85%  │
├──────────────────────────────────────────┤
│              ANIMUS QUORUM                │
│  Intent Graph · Consensus · Stigmergy     │
│  import convergent        906 tests  97%  │
└──────────────────────────────────────────┘
         + Bootstrap (daemon, wizard, dashboard)

Core — User-facing exocortex. Persistent memory (episodic, semantic, procedural via ChromaDB), identity system, 40+ CLI commands, integrations (Google Calendar, Todoist, filesystem, webhooks), and a cognitive layer supporting Anthropic, OpenAI, and Ollama with native tool use.

Forge — Headless orchestration engine. Declarative YAML workflows, 10 agent archetypes, per-agent token budgets, quality gates, SQLite checkpoint/resume, FastAPI backend, streaming execution logs, MCP tool execution, eval framework, and consensus voting. Deployed as a systemd service.

Quorum — Decentralized coordination protocol. Agents read a shared intent graph and self-adjust to be compatible with high-stability commitments — no inter-agent messaging, no supervisor bottleneck. Includes triumvirate voting, stigmergy, flocking behaviors, and phi-weighted stability scoring. Optional Rust PyO3 backend for performance. Published on PyPI as convergentAI.

Bootstrap — Install daemon, setup wizard, and ops dashboard for deploying Animus on new machines.


Usage

Define a workflow in YAML

# workflows/code-review.yml
name: code-review
agents:
  - role: researcher
    model: claude-sonnet-4-20250514
    budget: 4000
    task: "Analyze the codebase structure and identify patterns"
  - role: reviewer
    model: claude-sonnet-4-20250514
    budget: 8000
    task: "Review code for correctness, security, and maintainability"
    depends_on: [researcher]
gates:
  - after: reviewer
    check: quality_score >= 0.8

Run it

# Via CLI
cd packages/forge
animus-forge run workflows/code-review.yml

# Via API (Forge runs as a systemd service on port 8000)
curl -X POST http://localhost:8000/api/v1/workflows/run \
  -H "Content-Type: application/json" \
  -d '{"workflow": "code-review", "params": {"target": "./src"}}'

Use Quorum for agent coordination

from convergent import Intent, IntentGraph

graph = IntentGraph()

# Two agents register their intents
graph.register(Intent(
    agent_id="researcher",
    action="analyze",
    provides=["codebase_map"],
    stability=0.9
))
graph.register(Intent(
    agent_id="reviewer",
    action="review",
    requires=["codebase_map"],
    stability=0.7
))

# Find conflicts and resolve them without a supervisor
overlaps = graph.find_overlapping(agent_id="reviewer")

Repository Structure

animus/
├── packages/
│   ├── core/                    # import animus
│   │   ├── animus/              # Identity, memory, cognitive, CLI, integrations
│   │   ├── configs/             # YAML configs
│   │   └── tests/               # 1,736 tests
│   ├── forge/                   # import animus_forge
│   │   ├── src/animus_forge/    # Executor, agents, API, CLI, TUI, dashboard
│   │   ├── migrations/          # 14 SQL migrations
│   │   ├── workflows/           # YAML workflow definitions
│   │   ├── skills/              # Agent skill definitions
│   │   └── tests/               # 6,731 tests
│   ├── quorum/                  # import convergent (PyPI: convergentAI)
│   │   ├── python/convergent/   # Intent graph, voting, stigmergy, bridge
│   │   ├── src/                 # Rust PyO3 (optional performance layer)
│   │   └── tests/               # 906 tests
│   └── bootstrap/               # import animus_bootstrap
│       ├── src/animus_bootstrap/ # Daemon, wizard, dashboard
│       └── tests/               # 287 tests
├── docs/whitepapers/            # Architecture whitepapers
└── .github/workflows/           # CI: lint, test (per-package), security, CodeQL

Design Principles

Budget-first execution. Every agent has a token budget. Every workflow has a cost ceiling. Inspired by Toyota Production System — make cost visible, make waste impossible to ignore.

No supervisor bottleneck. The industry default for multi-agent coordination is a supervisor that watches everything. This burns tokens on monitoring and creates a single point of failure. Quorum replaces this with environmental awareness — agents observe shared state and independently converge, the way flocking birds coordinate without a lead bird.

Checkpoint/resume. All Forge workflows persist state to SQLite. If a pipeline fails at step 4 of 6, it restarts at step 4. No wasted compute.

Provider-agnostic. LLM calls go through a shared interface. Swap Claude for OpenAI or Ollama without touching agent code. Native tool use dispatches by provider.

Local-first. Your memory, your identity, your hardware. Nothing leaves unless you configure it to.


Tech Stack

Component Technology
Language Python 3.10+ (Core), Python 3.12+ (Forge), Rust (Quorum optional)
State SQLite WAL (checkpoints, budgets, migrations), ChromaDB (semantic memory)
Workflows Declarative YAML
LLM Claude API, OpenAI, Ollama (native tool use per provider)
API FastAPI (Forge backend, deployed as systemd service)
Coordination PyPI: convergentAI

Status

Package Version Tests Coverage Stage
Core 2.0.0 1,736 95% Active — CLI, memory, integrations
Forge 2.0.0 6,731 85% Active — deployed as systemd service
Quorum 1.1.0 906 97% Active — live on PyPI
Bootstrap 0.1.0 287 93% Active — daemon + wizard

Total: 9,660+ tests across 4 packages.


Background

This project grew out of 17+ years of enterprise operations experience, including applying Toyota Production System principles to AI workflow systematization. The orchestration layer (Forge) treats multi-agent execution the way a lean manufacturing line treats production — visible budgets, quality gates at every stage, and waste elimination through checkpoint/resume.

The coordination layer (Quorum) draws from biological systems research. Traditional multi-agent coordination uses either sequential execution (safe but slow) or supervisor patterns (flexible but expensive). Quorum introduces a third option: stigmergic coordination, where agents self-organize through shared environmental state rather than direct communication.


License

MIT

About

Personal AI exocortex — multi-agent orchestration + coordination protocol. Monorepo: Core, Forge (orchestration), Quorum (coordination). 9,300+ tests.

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Packages

No packages published

Contributors 3

  •  
  •  
  •