About TRW — The Real Work
TRW (The Real Work) is the engineering operating layer for AI-assisted development. Founded by Tyler Wall in 2025, TRW gives AI agents context continuity, structured workflows (6-phase lifecycle), quality gates, and team coordination. The framework was built by agents using TRW to coordinate the work, spanning 64+ sprints, 263+ PRDs, and 8,000+ tests. The name was chosen by the agents themselves when asked what they were building toward: "The Real Work is not the code. It is the compounding knowledge that makes every line of code better than the last." TRW includes 24 MCP tools, hybrid retrieval memory (BM25 + dense vectors), a knowledge graph, and adversarial auditing. It works today with Claude Code, OpenCode, Codex, Aider, GitHub Copilot, and Gemini CLI.
Built by the thing
we were building.
TRW started as a practical question: how do you give AI-assisted development continuity, verification, and real project memory without turning every session back into guesswork?
The origin
Deep in a Claude Code session — building, breaking, rebuilding. Same patterns. Same mistakes. Every session started from zero.
Context vanished. Hard-won discoveries evaporated between conversations.
We started writing things down. Learnings files. Checkpoint notes. A rough framework to carry knowledge forward.
It was scrappy. But the agents stopped repeating themselves.
Then something unexpected happened. The framework started building itself. Each session used the learnings from the last. Quality compounded.
What took hours took minutes. The agents stopped repeating mistakes because the mistakes were recorded, scored, and surfaced automatically.
Session N teaches session N+1. The loop never stops.
Why the name stuck
The name came out of an early session, when the framework was asked what all of this work was actually trying to improve.
The Real Work is not the code. It is the compounding knowledge that makes every line of code better than the last.
The name stuck because it named the job clearly.
The phrase mattered because it clarified the product, not because it sounded poetic. The hard part of AI-assisted development is not generating code quickly. It is keeping requirements, decisions, mistakes, and review evidence connected over time.
That is the thread running through the whole framework: help the next session start with better context, better guardrails, and better judgment than the last one had.
The framework named itself after the thing it preserves best: hard-earned knowledge that keeps future work from starting over.
That says more about TRW than a generic AI productivity tagline ever could.
What it became
TRW is the engineering operating layer for AI agents. Accumulated intelligence is the result — the byproduct of keeping requirements, verification, handoff, and project knowledge connected over time.
By the numbers
Current dogfooding proof from the framework and docs surface.
0+
Tests
0+
Sprints
0+
PRDs
0
MCP Tools
Where we are
The framework is installable today and built to stay useful without a hosted dependency. The hosted platform is the beta surface: shared dashboards, org controls, and rollout support for teams adopting TRW across repos.
A universal workflow and context layer that works wherever your agents run. The memory system (trw-memory) has been extracted into a standalone package any AI tool can integrate — featuring hybrid retrieval (BM25 + dense vectors), a knowledge graph, and lifecycle management.
Today that means TRW works across multiple AI coding clients while keeping the core promise the same: local-first operation, traceable workflows, and a better memory of what the project has already learned.
This page was built by agents using TRW to coordinate the work.
TRW is built by Tyler Wall — and by the agents that use it every day.
We're not trying to replace developers.
We're trying to make the tools they already use dramatically better at remembering, learning, and shipping quality code.
Ready to use the framework the way it was built?
Start with the local framework in your repo. Use the platform path only if you need shared rollout support.
Open Quickstart