Do I need to modify my codebase?

No. TRW adds a single configuration entry to your AI coding tool. Your source code, repository structure, and CI pipeline remain completely untouched. TRW stores its knowledge in a small .trw/ directory — that’s the only addition.

Can multiple AI agents work together?

Yes. TRW supports focused roles such as researcher, implementer, reviewer, auditor, and PRD groomer. They share tasks through structured messaging, keep ownership boundaries clear, and pass through verification before delivery.

Project state stays repo-local by default. Optional telemetry or hosted sync requires explicit configuration, and the hosted path is currently gated while onboarding hardens.

What does ‘Accumulated Intelligence’ mean?

Accumulated Intelligence is the outcome, not the product category. Each session leaves behind reusable project knowledge that gets scored, preserved, and resurfaced when the next task needs it.

How is this different from Mem0 or native Claude Code memory?

Memory is table stakes. TRW connects memory to requirements, orchestration, verification, and handoff. The differentiator is not that something gets stored — it is that stored knowledge changes how work gets planned, reviewed, validated, and resumed.

How much does it cost?

TRW is free during the current gated beta. We’ll introduce paid tiers for teams and enterprise features after launch, but a generous free tier will always exist for individual developers.

What happens to my data if I stop using TRW?

Remove the config entry and TRW disappears. The .trw/ directory contains plain YAML and Markdown — human-readable, version-controlled, and yours to keep or delete. There is zero lock-in by design.

Does it work with my existing tools?

TRW works with Claude Code, OpenCode, Codex, and Aider today, alongside editor surfaces like VS Code, Windsurf, Zed, and JetBrains. GitHub Copilot and Gemini CLI are currently in beta, and Cursor remains on the roadmap.

Engineering layerEngineering operating layer for AI agents

Your AI can write code. TRW makes it shippable.

TRW sits above your model, IDE, and memory. It keeps requirements, execution, verification, handoff, and preserved project knowledge connected from one session to the next.

Get Early Access See the workflow

Knowledge base

490learnings

Database migrations must run before deployment, not after

deploymentdeploy-pipeline

critical

0.95

JWT refresh tokens need rotation on every use

securityauth-service

critical

0.93

React useEffect cleanup prevents memory leaks on unmount

frontendplatform-ui

high

0.85

Durable project context · local-first

Knowledge base

487learnings

Signal from 249 sessions · recalled at every session start

Install the engineering operating layer your AI tools are missing

Read the quickstart · Request beta access when TRW fits your workflow

Read Quickstart

Works with your tools

Beta

Claude Code

VS Code

Windsurf

Beta

GitHub Copilot

Beta

Gemini CLI

Soon

Cursor

Zed

JetBrains

Aider

Beta

OpenCode

Beta

Codex

LangChain

LlamaIndex

CrewAI

OpenAI

Beta

Claude Code

VS Code

Windsurf

Beta

GitHub Copilot

Beta

Gemini CLI

Soon

Cursor

Zed

JetBrains

Aider

Beta

OpenCode

Beta

Codex

LangChain

LlamaIndex

CrewAI

OpenAI

Knowledge continuity

Every session recalls what the last one learned.

TRW is designed so that session 50 can start from richer retained context than session 1. Whether that translates to measurable capability lift is an open empirical question — see /metrics for current evidence.

Every other AI tool starts from scratch. TRW preserves state in .trw/ and replays it at session start.

Context surfaced per session

—recalls

accumulated: 0 · indicative values

Session 1

Baseline

Session 10

Learns your codebase

Session 25

Learns your team

Session 50

Learns your domain

Session 100+

Compounds

See how 3000+ learnings are recalled across sessions

Governed execution

A workflow that keeps specs, code, and verification connected

Research, plan, implement, validate, review, and deliver — with evidence at each handoff and reusable knowledge at the end.

Research

Plan

Implement

Validate

Review

Deliver

Research

Recall prior work & inspect the codebase

Plan

Choose the approach & map dependencies

Implement

Ship with checkpoints so work survives compaction

Validate

Run tests, type checks & coverage gates

Review

Audit correctness, quality & security

Deliver

Publish artifacts & persist learnings

Run complete · learnings signed

All phases signed

What's Inside

Six capabilities that turn AI output into engineering work

Memory is one piece of the system. Requirements, verification, handoffs, and coordination are what make the work trustworthy.

Layer 1 of 6

Requirements Engineering

AI stops building the wrong thing

Turns vague requests into structured specs with acceptance criteria — agents verify work against the spec, not vibes.

See requirements engineering in motion

What's Inside

Six capabilities that turn AI output into engineering work

Memory is one piece of the system. Requirements, verification, handoffs, and coordination are what make the work trustworthy.

Requirements Engineering

AI stops building the wrong thing

Turns vague requests into structured specs with acceptance criteria — agents verify work against the spec, not vibes.

Persistent Knowledge

Your agents remember everything

Every pattern, gotcha, and decision persists across sessions with impact scoring that surfaces the most valuable knowledge automatically.

Cross-Project Intelligence

What one project learns, all projects know

A discovery in one project automatically surfaces when agents encounter similar problems elsewhere.

Automated Verification

Ship only code that actually works

Tests, types, and coverage enforced automatically before anything ships — works with your existing suite.

Team Coordination

Multiple agents, one coherent output

Specialized roles — researcher, implementer, reviewer, auditor, and PRD groomer — each with focused context and clear ownership boundaries.

Structured Workflows

Consistent quality without constant oversight

Every task follows a defined lifecycle with quality gates between phases and checkpoints that preserve progress.

Proof

Proof from a codebase that runs on TRW

Tracked repo metrics from work reviewed and shipped with TRW itself.

Patterns Discovered

3,000

Gotchas, patterns, and decisions your agents never have to rediscover.

Agent Actions Tracked

Tool calls, checkpoints, and deliveries captured across TRW runs in this repo.

Projects Using TRW

Telemetry-backed installs when available, with this repo as the baseline snapshot.

Quality Compliance

How consistently agents follow the engineering workflow before shipping.

Live telemetry when available · Otherwise a curated snapshot from this repo·View all metrics

100+

Production hardening, not just prototypes

Rate limiting, audit logging, and input validation were implemented and reviewed in one coordinated TRW-guided pass.

3000+

Learnings captured from TRW sessions

When session 1 discovers a gotcha, it is stored in `.trw/` and recalled by later sessions. Whether this yields measurable capability lift is actively measured.

15000+

Quality gates, not wishful thinking

15000+ tests, strict type checking, and independent code review — all enforced automatically before code ships.

Read Quickstart Join Beta

Review the workflow first. Request beta access when you are ready to try it.

Why it stays in the repo

Work that compounds instead of resetting

Each task leaves behind usable project knowledge — not just output that disappears when the session ends.

Local · private

Zero data risk

Everything runs locally — no cloud dependency, no data exfiltration, remove anytime.

Durable · transferable

Knowledge that outlasts your team

When engineers leave, their AI agent's accumulated knowledge transfers instantly to the next person.

Compounding · cross-project

Compound returns at scale

One agent's discovery benefits every project — a genuine compounding asset.

Install

Install once. Improve the next session.

Once approved, TRW wires in as a single configuration entry. No app code changes. No migrations. Your existing tools stay in place. Remove it anytime — zero lock-in.

TTY_INSTALLone command

1	# Install or upgrade TRW in your project
2	curl -fsSL https://trwframework.com/install.sh \| bash

Request beta access to continue the authenticated install

Offline by default

Nothing leaves your machine. Telemetry is opt-in.

Zero lock-in

Remove the config entry and TRW disappears. Your code stays yours.

Language-agnostic

TRW works with any codebase. Verification hooks can run your existing test stack, not a new one.

FAQ

Common questions

AI writes code. TRW makes it engineering.

Install in two minutes. Keep requirements, verification, and learnings connected from one session to the next.

Read Quickstart Join Beta

Access is currently limited while onboarding hardens · Start with the docs · Zero lock-in

Your AI can write code. TRW makes it verifiable.shippable.

Every session recalls what the last one learned.

A workflow that keeps specs, code, and verification connected

Six capabilities that turn AI output into engineering work

Requirements Engineering

Six capabilities that turn AI output into engineering work

Requirements Engineering

Persistent Knowledge

Cross-Project Intelligence

Automated Verification

Team Coordination

Structured Workflows

Proof from a codebase that runs on TRW

Work that compounds instead of resetting

Zero data risk

Knowledge that outlasts your team

Compound returns at scale

Install once. Improve the next session.

Common questions

>What is TRW?

>Do I need to modify my codebase?

>Can multiple AI agents work together?

>Is my data private?

>What does ‘Accumulated Intelligence’ mean?

>How is this different from Mem0 or native Claude Code memory?

>How much does it cost?

>What happens to my data if I stop using TRW?

>Does it work with my existing tools?

AI writes code. TRW makes it engineering.

Your AI can write code. TRW makes it shippable.