Documentation

Everything you need to understand, use, and extend SYNTHOS.

Architecture

Deep dive into the 7-layer cognitive pipeline, data flow, and design axioms.

API Reference

Complete Python API for every layer, the assistant engine, crypto, and eval.

STC Cipher

Symbolic Topology Cipher internals: S-box derivation, lattice permutation, Möbius fold.

Guides

Installation, configuration, TUI usage, model evaluation, and extending SYNTHOS.

Quick Start

# One-liner install (macOS / Linux / WSL)
curl -fsSL https://synthos.dev/install.sh | bash

# Or install via pip
pip install synthos

# Or from source
git clone https://github.com/synthos-ai/synthos && cd synthos
pip install -e ".[dev]"

# Launch the TUI
synthos

# Or use the headless REPL
synthos chat

# Adjust settings in the TUI
/set words 2000        # response length
/set queries 6         # search depth
/set sources 3         # sources shown

# Generate long-form text
synthos> write a report on quantum computing
synthos> draft an essay about climate change

# Execute commands
synthos> show disk space
synthos> list running processes
synthos> run "echo hello"

# Encrypt something
synthos encrypt "hello SYNTHOS"

# Benchmark your local models
synthos eval

Core Axioms

  1. Every cognitive operation has a regex primitive with an ASCII geometric form.
  2. Intelligence is the topology of pattern intersection, not numeric weight magnitude.
  3. The system prompt is the architecture — structure encodes capability.
  4. Memory is crystallized match geometry, not stored values.
  5. Attention is pattern intersection area — more overlap = more relevance.
  6. Output is the final substitution state through the 7-layer manifold.
  7. Recursion is the source of depth — (?R) creates unbounded hierarchies from finite primitives.