About Mission
An AI systems lab. A continuity-first workflow.
ThinkCore is the public identity and reference for how I build — the stack, the workflow, the decisions, and the principles that shape every project.
AI-native work is fragmented across chats, files, repos, tools, and time. ThinkCore exists to make that evolution coherent — and to make the ecosystem legible from the outside.
About Problem
AI-native creators lose continuity as work spreads across disconnected environments.
Project context often begins in a chat, moves into files, becomes distributed across repositories, changes through tool output, and is later reconstructed from memory. Decisions may be made quickly, revised later, or buried inside long conversations.
The result: project evolution that becomes hard to inspect, trust, or continue.
- Chat context is separated from the files it influenced
- Decisions are made without a durable record of rationale
- Repository changes lack connection to earlier planning
- Architecture shifts are not tracked as the project changes
- Source material is separated from conclusions derived from it
- Workflow state is recreated manually from scattered notes
- Chronology is unclear after multiple sessions or collaborators
- Important context becomes difficult to retrieve when needed
Principles Philosophy
Six principles that shape every decision.
Continuity First
The primary value is coherent project evolution over time. Every structure should help future readers understand what changed, why it changed, and how current state was reached.
Practical Organization
Information is organized by use, meaning, and retrievability. Folders and documents should make project context easier to locate without requiring specialized knowledge or hidden conventions.
Source Awareness
Derived context remains connected to the sources that shaped it. Summaries, decisions, and architecture notes are more useful when their origin is clear.
Human Governance
Humans remain responsible for project direction, acceptance, and interpretation. Tools help organize and retrieve context — they do not blur responsibility for decisions.
Restrained Architecture
Architecture grows from demonstrated workflow needs. Abstractions, categories, and processes are not added before they make continuity clearer.
Observable Change
Changes to project understanding should be inspectable. Records show sequence, rationale, and relationship to existing work where practical.
Principles Decisions
Architectural posture and the reasoning behind it.
Documentation before runtime
Structure and documentation are established before any runtime systems are introduced. This prevents architecture from being shaped by implementation urgency rather than actual need.
Semantic organization over storage location
Information is organized by meaning and use — not by file type, date, or storage system. A decision and the conversation that produced it belong near each other.
Restrained growth
New layers are only added when they reduce confusion, preserve meaning, or make governance clearer. Three similar structures are better than a premature abstraction.
Explicit layer boundaries
ThinkCore, Liahona, CCC, and KINDEX have distinct roles. Operational command language belongs in CCC. Reflective orientation belongs in Liahona. Each layer preserves its own character.
No surveillance, no total capture
The focus is operational continuity for project and workflow work. ThinkCore does not attempt to capture all human context — only the continuity that makes reconstructable knowledge work possible.
Chats are temporary. Records are not.
Operational continuity must survive beyond context windows, AI sessions, tools, deployments, workflow changes, and model providers. The record outlasts any single tool.
System Ecosystem
Four distinct layers. Each with a specific role.
ThinkCore
The ecosystem identity, routing layer, and public home for AI systems work. Makes the lab's research, projects, and system boundaries legible. Does not absorb the jobs of sibling layers.
Liahona.AI
The reflective orientation layer. Quiet, symbolic, continuity-aware, and artifact-centered. Not productivity software. Not an operational dashboard. A different surface entirely.
CCC
The operational command layer. Tactical, diagnostic, telemetry-aware, and workflow-focused. Where command-center language and live system state belong.
KINDEX
Human continuity and archive context. The long-term record layer outside the operational surface.
System Stack
The operational environments surrounding AI-native work.
- Claude Anthropic
- GPT-4 / GPT-5 OpenAI
- Codex OpenAI
- Cursor
- VS Code
- GitHub
- Node.js
- PM2
- Docker
- PostgreSQL
- SQLite
- Render
- GitHub Pages
- Vercel
- Discord
- Obsidian
- Apple Notes
- MCP Model Context Protocol
- HTML / CSS / JS Vanilla, no framework
System Deployment
Two surfaces. One for previewing, one for production.
Vercel
Every branch push automatically gets a unique preview URL. Use this to review changes before merging — no local server needed.
- Push branch to GitHub
- Vercel detects the push and deploys
- Preview URL appears in the PR / Vercel dashboard
- Review, iterate, merge when ready
GitHub Pages
The live site at thinkcore.io. Deploys automatically from the main branch within a minute of merging.
- Merge branch → main
- GitHub Pages builds and deploys
- thinkcore.io reflects the change
vercel.json at the repo root marks this as a plain HTML/CSS/JS site with no build step, so Vercel skips framework detection.
Reference Workflow
The goal is not to replace existing tools — it is to preserve the relationships between them.
-
1
Capture — Relevant project context from chats, files, decisions, and source material
-
2
Organize — Clear project areas by use and meaning, not just storage location
-
3
Preserve Chronology — When sequence affects understanding, order matters
-
4
Connect Decisions — To rationale and source material, not just outcomes
-
5
Track Architecture — Changes as the project evolves, not retrospectively
-
6
Keep Retrievable — For future human and AI-assisted work
- What changed?
- Why did it change?
- What source material influenced the change?
- What decision or constraint shaped the work?
- What remains unresolved?
- Where should future work continue?
Reference Vocabulary
Shared terms. Precise definitions prevent drift.
- Continuity
- The preservation of coherent project evolution across chats, tools, files, repositories, decisions, workflows, sources, architecture, chronology, and time.
- Continuity Entry
- The primary record of a meaningful operational event in a project. May describe a decision, implementation, workflow change, architecture shift, source update, deployment state, bug, fix, discovery, or other event that affects project evolution.
- Environment
- The operational context surrounding work — AI tools, apps, communication systems, repositories, terminals, local files, operating systems, deployment surfaces, and other places where project state exists.
- Source
- Evidence or material that supports continuity context — conversations, files, commits, documents, logs, screenshots, URLs, notes, exports, architecture records.
- Relationship
- A typed connection between continuity records — connecting entries to sources, decisions, environments, architecture context, workflow state, outcomes, later work, or related entries.
- Operational Lineage
- The sequence of meaningful project work across systems and time. Helps reconstruct what happened, why it happened, which environments were involved, what sources support the record, and how later work depends on earlier work.
- Workflow Adapter
- A bridge between ThinkCore and an external workflow that preserves continuity across a tool or environment without replacing the user's workflow.
- Decision
- A recorded choice that includes enough context to understand its rationale, constraints, source material, and effect on later work.
- Provenance
- The traceable origin of information — where context came from, what source supports it, and how it entered the project record.
- AI-Native Creation
- Creation work that uses AI systems as active collaborators across planning, writing, design, code, analysis, review, and iteration.
- Project State
- The current understood condition of a project across its fragmented systems, records, decisions, sources, and environments.
- Semantic Organization
- The arrangement of information by meaning and use rather than only by storage location or file type.