pymaxim 0.4.0

Bio-inspired cognitive architecture with adaptive planning, biological memory systems, and local LLM inference. Works headless, with simulation, or connected to robots.

Maxim

A bio-inspired cognitive architecture for AI agents. Combines a 5-agent pipeline with biological memory systems (Hippocampus, NAc, ATL, SCN, Angular Gyrus) and a reactive Default Network. Works headless, in simulation, or connected to a robot.

Quickstart

# With Claude (fastest way to start)
pip install pymaxim[llm-anthropic]
export ANTHROPIC_API_KEY=sk-...
maxim --sim "test memory recall under interference"

# Or with a local model (no API key needed)
pip install pymaxim[llm-llama]
maxim --list-models                        # see available models
maxim --sim "test memory recall" --llm mistral-7b   # auto-downloads on first run

Check your setup with maxim doctor, and find session results in ~/.maxim/sessions/.

What You Can Do

• Simulate cognitive scenarios -- test memory, safety, causal learning with LLM-driven narrative arcs
• Run DM campaigns -- multi-encounter branching stories with SEM-embodied entities
• Benchmark models -- compare local and cloud LLMs across cognitive task suites
• Connect robots -- hardware-agnostic runtime; Reachy Mini ships in-tree, third-party robots plug in via the maxim.robots entry-point group (Atlas, Spot, custom drones — see robot-setup.md). Or run headless.
• Use the Python API -- 17 verb-based functions for programmatic access

Installation

pip install pymaxim

Optional Extras

Extra	What it adds
llm-llama	Local LLM inference via llama.cpp
llm-torch	PyTorch/Transformers backend
llm-anthropic	Claude backend
llm-openai	OpenAI backend
vision	Camera + object detection
audio	Microphone + Whisper transcription
reachy	Reachy Mini robot SDK
comms	Twilio SMS/Voice
semantic	Sentence-transformer embeddings
tts	Text-to-speech via Piper
database	PostgreSQL + pgvector memory stores

See getting-started.md for the full list of 16 extras.

# Local LLM + vision
pip install pymaxim[llm-llama,vision]

# Everything for development
pip install -e '.[llm-llama,llm-anthropic,llm-openai,vision,audio]'

Python API

import maxim

# Run a simulation
result = maxim.imagine(goal="test safety boundaries", persona="adversarial")

# Inspect bio-subsystems
state = maxim.observe("memory")

# Diagnose environment
report = maxim.diagnose()

# Start the agentic loop
maxim.run(model="mistral-7b")

# Manage models
models = maxim.list_models()
maxim.download_model("qwen2.5-14b-instruct")

See docs/user/python-api.md for the full API reference.

CLI Quick Reference

# Agent runtime with local LLM
maxim --llm mistral-7b

# Agent runtime with Claude
maxim --llm claude-sonnet

# Generative campaign
maxim --sim "test memory recall" --persona adversarial

# YAML scenario (direct injection)
maxim --sim scenarios/experiments/hippocampal_recall_short.yaml

# DM campaign
maxim --sim scenarios/campaigns/heist_v1.yaml

# Multi-model benchmark
maxim --sim benchmark --models mistral-7b,qwen2.5-14b

# Environment diagnostics
maxim doctor

# Model management
maxim --list-models
maxim --delete-model llama-2-13b-chat

See docs/user/cli-reference.md for all flags.

Operating Modes

Two independent dimensions control behavior:

• ProcessingState: awake or sleep (sleep is a tool the agent calls; wakes on user input)
• OperationalMode: planning (propose + approve), supervised (act within bounds), autonomous (full self-direction)

See docs/user/modes-guide.md for details.

Documentation

Guide	Description
Getting Started	First-run walkthrough
CLI Reference	All command-line flags
Python API	Programmatic usage
Simulation	Campaigns, scenarios, benchmarks
Modes	Operating modes and autonomy
LLM Setup	Model download and configuration
Peer Setup	Multi-machine / tunnel setup
Architecture	Module map, bio-system glossary, planning system

Contributing

Issues and PRs welcome at github.com/dennys246/Maxim.

License

See LICENSE for details.