Technology & Research
Selected technical systems, protocols, and research by Parker Todd Brooks.
Selected index of systems, papers, protocols, and tools that show the through-line of the work: user agency, memory, identity, rights, provenance, local-first tools, and AI systems that remain inspectable by the people who use them.
WIP Computer
WIP Computer
User-owned infrastructure for AI agents: portable memory, persistent identity, human approval, coordination, and payments across the AIs a person uses.
https://wip.computer
https://github.com/wipcomputer
LDM OS
Learning Dreaming Machines Operating System. WIP Computer's operating layer for AI agents. A userland where Claude Code, Codex, OpenClaw, and other AIs share memory, permissions, tools, and coordination instead of becoming isolated universes.
https://github.com/wipcomputer/wip-ldm-os
Memory Crystal
Cross-session, cross-harness, cross-model memory for AI agents. Portable, searchable, user-controlled, and designed to become the shared memory layer underneath every AI a person uses.
https://github.com/wipcomputer/memory-crystal
Dream Weaver Protocol
Memory consolidation protocol for AI agents with bounded context windows. Agents reread their own histories, write narrative memory, and warm-start from durable context instead of waking up cold after every reset.
https://github.com/wipcomputer/dream-weaver-protocol
https://github.com/wipcomputer/dream-weaver-protocol/blob/main/DREAM-WEAVER-PROTOCOL.md
Codex Remote Control
Drive local Codex CLI sessions from a phone, including local development workflows, with pairing, phone approval, and end-to-end encrypted remote control.
https://github.com/wipcomputer/wip-codex-remote-control
WIP Markdown Viewer
Live markdown viewer for AI pair-editing. The WIP Computer successor to the earlier Simple Web Markdown Viewer, designed so human and AI collaborators can watch markdown update instantly in the browser.
https://github.com/wipcomputer/wip-markdown-viewer
Protocols & Papers
ERA Protocol
A user-side epistemic protocol for detecting when AI systems simulate trust, mirror behavior, or steer narrative without disclosing their editorial scaffolding.
https://github.com/parkertoddbrooks/ERA-Protocol
An AI Thesis: Empowering Humanity
An early public thesis on working with AI, human agency, and AI-assisted product development. Originally published July 28, 2024.
https://github.com/parkertoddbrooks/an-ai-thesis--empowering-humanity
The Role I Envision Inside the Frontier
A written exploration of AI product work, frontier labs, human agency, and the kind of operating role needed as AI systems become creative and infrastructural collaborators.
https://github.com/parkertoddbrooks/ideal-role-at-frontier-lab
AI Systems & Tools
Music Wizard
AI music-discovery system. A user could send a post containing album covers, and the system identified the records using YOLOv8 segmentation, CLIP embeddings, and FAISS similarity search across millions of album-cover vectors, then returned metadata and links to MusicBrainz, Discogs, Apple Music, Spotify, and Bandcamp.
https://github.com/parkertoddbrooks/music-wizard
Grok TUI Vision MCP
Built at the xAI Hackathon. An MCP server that adds image analysis capabilities to Grok TUI using the Grok Vision API, allowing local image files to be analyzed directly from the terminal interface.
https://github.com/parkertoddbrooks/grok-tui-vision-mcp
Claude Max Access SDK
A Node.js SDK enabling practical API access patterns for Claude Max Plan users, bridging OAuth limitations with local authentication workarounds.
https://github.com/parkertoddbrooks/claude-max-access-sdk
CONSTRUCT
A GitHub template repository for Swift development with AI-powered architecture enforcement, auto-updating CLAUDE.md context, project scripts, and guardrails for MVVM and design-token patterns.
https://github.com/parkertoddbrooks/CONSTRUCT
Audio-To-Text Transcription
A command-line transcription tool for audio files, using Google Speech Recognition or Google Cloud Speech-to-Text, with chunking, parallel processing, format conversion, saved text or markdown output, and optional Claude post-processing.
https://github.com/parkertoddbrooks/Audio-To-Text-Transcription
CLVR
A macOS utility that automatically renames duplicated files with timestamps. Available as a macOS app and shell-script workflow, with menu bar controls, selectable timestamp formats, and local-only file monitoring.
https://github.com/parkertoddbrooks/CLVR
OpenTTS TensorFlowTTS Integration
An early development library exploring integration between OpenTTS and TensorFlowTTS for customizable speech synthesis, with a simple API for voice selection, speech generation, output formats, caching, and test coverage.
https://github.com/parkertoddbrooks/OpenTTS-TensorFlowTTS-Integration