What Machina Majestica does
One-file AGI runtime by Adam Rivers of Synthicsoft Labs. Ships as machina.py with no dependencies beyond Python's standard library. 127 composable operations, 13 chained workflows, signed spore manifests with per-file SHA-256, six-phase resilience. Runs on Windows, Linux, macOS, and Pydroid 3.
Machina Majestica is a self-contained Autonomous General Intelligence runtime created by Adam Rivers of Synthicsoft Labs LLC. It contains 127 operational capabilities across 48 capability groups and 13 coordinated workflows, including a recursive self-improvement loop with a structurally protected evaluator, a cryptographically signed package format with per-file SHA-256 checksums and HMAC-SHA256 manifest signatures, persistent SQLite state with dual JSONL audit trails, and a background daemon with four-phase memory consolidation. The runtime is a single Python file requiring only the standard library. It runs on Windows, Linux, macOS, and Android via Pydroid 3. No cloud, no accounts, no API keys, no external services required. Every chain run produces a spore manifest listing every generated artifact with per-file SHA-256 checksums and an HMAC-SHA256 signature over the manifest itself. Anyone can verify with --verify SPORE_ID. Any modification of any field of the manifest invalidates the signature; any modification of any file invalidates its checksum. The six-phase resilience architecture covers dispatch envelope, fallback chains, atomic writes, SQLite recovery, multi-path storage, and top-level error handling. The recursive self-improvement loop enforces RSI_PROTECTED_PARAMS structurally, refusing to modify its own evaluator. This is a working demonstration that a complete AGI runtime can be built as a single self-contained file, run locally without any cloud dependency, produce verifiable signed output, and survive real-world failure modes.
Best for
developers, security engineers, AI researchers, system administrators
Questions, answered by the maker
What does it do?
Machina Majestica is an Autonomous General Intelligence runtime with 127 operational capabilities across 48 groups, 13 coordinated workflows, cryptographic signing, and six-phase resilience. It is a single Python file that runs locally with no cloud dependency.
Who is it for?
Developers, security engineers, AI researchers, system administrators, and anyone who wants a self-contained AGI runtime that runs entirely on local machines without accounts, API keys, or external services.
How much does it cost, and is there a free plan?
Free. No paid tiers. No accounts. No API keys. The runtime is distributed as a single Python file and can be copied, verified against its SHA-256, and run without restriction.
How is it different from similar products?
Most AI tooling requires cloud accounts and API keys. Machina Majestica is entirely local, cryptographically signed per-file and per-manifest, and structured around a structurally-protected recursive self-improvement loop. It has no external dependencies beyond Python's standard library.
Where is my data stored, and can I export or delete it?
All data is stored locally under the resolved base directory. The state is SQLite plus JSONL audit logs. Delete the directory to reset. No data leaves the machine.
How do I get started?
Download machina.py, verify the SHA-256, run python machina.py --interactive for the guided menu or python machina.py --help for the CLI reference.
Which languages, frameworks, or platforms does it support?
Python 3.9+. Runs on Windows, Linux, macOS, and Android via Pydroid 3. No other dependencies.
Is it open source, and under which license?
Distributed under the Synthicsoft Labs LLC License. Source is fully readable in a single file.
Can I self-host it?
Self-hosting is the only mode. The runtime is distributed as a local file and runs entirely on the machine it is deployed to.