What is Tenkr?

Not a Chatbot

You have heard the pitch: “an AI assistant that runs your work.” Here is what that actually means.

Tenkr is not a chatbot you ask questions to. A chatbot waits for your prompt, generates a response, and forgets. Tenkr is an AI operating system that turns a capable but amnesiac language model into a persistent, context-aware operator that knows your project, follows your methodology, and proves its work.

The model underneath is the same. The context engineering makes it behave like a different system entirely.

How Tenkr Differs from a Raw AI Model

If you have used an AI coding assistant before, Tenkr will feel different immediately.

Memory. A raw AI model forgets everything between sessions. Tenkr maintains multiple memory layers: daily session notes written automatically, searchable knowledge documents, persistent memory that accumulates decisions and patterns, and codebase profiles for external repos. Session 50 is smarter than session 1 because the context compounds.

Methodology. A raw AI model approaches every task ad hoc. Tenkr has methodology skills that fire at specific moments: brainstorming before creative work, validation-first execution before producing deliverables, verification-before-completion before claiming anything is done, systematic debugging when things break. These are not suggestions the model might remember — they are injected into context at the right moment by hooks, so the model sees them exactly when it needs them.

Hooks. A raw AI model has no lifecycle awareness. Tenkr hooks into lifecycle events at key moments: session start, prompt submission, tool use, and session end. Each event is a chance to inject context, modify behavior, or enforce quality gates. These interventions happen invisibly — you see the results, not the mechanism. For a deep dive, see our article on How Context Engineering Works.

Skills and agents. A raw AI model is one general-purpose assistant. Tenkr has a library of skills (atomic procedures for specific tasks) and agents (domain experts with baked-in methodology). A data analyst agent knows how to query databases with cost guardrails. A workflow automation agent knows how to build integrations node by node. Each brings domain knowledge that a raw model would need to rediscover every session.

Commands. Beyond natural language, Tenkr responds to structured commands that invoke specific workflows. Triage your inbox. Query a database in plain English. Start a full application build pipeline. Commands are the entry points — thin triggers that route to the right agent with the right context.

The Three Domains

Tenkr organizes everything into three domains with strict boundaries.

Workshop is where you build things. The application building pipeline lives here — from raw project idea through requirements, wireframes, database design, and spec generation to a running application. If you are creating something new, you are in Workshop.

Methodology is where you plan, reason, and extend the system. The brainstorm-to-build pipeline handles complex multi-step work. Codebase profiling and cross-repo tasks live here too. If you are planning work or managing how work gets done, you are in Methodology.

Utility handles operational tools — searching your knowledge base, reindexing memory, managing pull requests. These are the maintenance and support functions that keep everything running.

Beyond these three, Tenkr has domain-specific modules for specialized work: Google Workspace operations, database querying, and workflow automations. Each follows the same architectural pattern but focuses on a specific integration domain.

What “Your Own Tenkr” Means

Tenkr is not a single product you use out of the box. It is a framework — an architecture of hooks, skills, agents, commands, patterns, and memory that can be personalized to any domain.

When you start with Tenkr, you get the universal framework: the hook system, the methodology skills, the command structure, the memory layer, and a set of domain-specific capabilities. This is the same foundation everyone starts with.

From there, your instance diverges. You profile your department's codebases, and Tenkr learns their architecture. You build custom skills for your team's repeating workflows. You create patterns that capture your domain's conventions. You add knowledge documents that get indexed and searched. Over weeks, your Tenkr accumulates context that makes it increasingly effective for your specific work.

Two users running the same Tenkr platform will have very different instances after a month. One might have deep expertise in financial reporting databases. Another might have profiled three frontend codebases and built custom patterns for their team's code review process. The framework is shared. The intelligence is personal.

This is the core idea: Tenkr is not a tool you use. It is a system that learns your domain, remembers your decisions, follows your methodology, and compounds over time. Week 8 is dramatically better than week 1 — not because the model improved, but because the context did.