aOa - Angle O(1)f Attack
Copyright 2025-2026 MVP-Scale.com
Author: Corey Gorman

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

---

Patent Notice

The one-bit embedding model and dimensional signal scoring system
implemented in this software are Patent Pending. This includes the
method of encoding code quality signals as single-bit positions within
a tiered bitmask vector, and the multi-signal scoring formula that
combines severity, density, co-occurrence, clustering, and breadth
signals from per-line bitmask topology.

Use of these methods within this licensed software is granted under
Apache License 2.0, Section 3 (Grant of Patent License). Use of these
methods outside this software may require a separate license.

---

A Word from the Founder

MVP-Scale.com is building the world's first SaaS foundry -- a community
and pipeline for launching foundational MVPs at 5-10x speed. This tool
is one piece of that pipeline.

aOa started because we watched Claude Code burn through tokens
rediscovering code it already found. Every session, the same searches.
The same files. The same wasted context. The fix isn't more AI -- it's
a map. aOa builds that map locally, silently, from the signals Claude
already produces.

The original aOa was a Python prototype that validated semantic code
intelligence for Claude Code. It worked -- but it needed Docker, Redis,
and Python just to run. The Go rewrite eliminates all of that. One binary.
No containers. No services. No runtime dependencies.

It went through a clean-room rewrite -- not porting Python line by line,
but rebuilding from behavioral specs and test fixtures. The result:
100,000x faster autotune, 16-30x faster search, 8x less memory, and a
single ~9MB binary that supports 509 languages via tree-sitter grammars
and 134 semantic domains embedded at compile time.

We're not building through vibe coding or AI hype. We're taking real
ideas, pushing them fast, iterating relentlessly, and shipping things
that work. This is our pipeline in practice.

If you build on this, fork it, extend it, or just find it useful --
we'd love to hear from you. We're building more tools like this, and
your use case might shape what comes next.

  https://github.com/MVP-Scale/aOa/discussions
  https://mvp-scale.com

This is the way.

-- Corey Gorman, Founder, MVP-Scale.com
