Start heresource checked

Contents and learning routes

The complete product-first route through T3 Code, followed by its control plane, six providers, work lifecycle, clients, device hosts, remote paths, and shipping system.

What this chapter resolves

Choose a product, workflow, or implementation route.
See every product and technical chapter in the guide.

The product guide comes first. The technical guide then follows one request from the client through authorization, durable state, provider execution, work services, client convergence, remote access, and release operations.

Start here

Cover — scope and source synchronization.
How to read this guide — product, workflow, and architecture routes.
Contents and learning routes — this complete map.

Part I · Product guide

T3 Code as a product
Install and reach a useful first thread
Environments, clients, and connection paths
Organize projects, threads, and worktrees
Compose tasks with the right context
Choose providers, models, accounts, and permissions
The workbench: files, terminals, browser preview, and SnapShots
Source control: checkpoints, pull requests, reviews, and stacks
Remote environments and unattended work
Mobile: supervise agents from anywhere
Device lab: simulators, emulators, and agent-driven testing
Personalize, measure usage, update, and protect privacy
Complete recipes and troubleshooting

Part II · Architecture orientation and boundaries

The complete system map
One request, every boundary
Control surface, not agent brain
Environment · project · thread · turn · session
Repository and dependency atlas
Runtime topologies and technology choices

Part III · Boot and connect

The npx t3 bootstrap path
Server composition, activation, and readiness
HTTP, WebSocket RPC, snapshots, and resume
Pairing, credentials, TTLs, and scopes

Part IV · The durable domain kernel

Commands, invariants, and the boundary of atomicity
Events, receipts, idempotency, and the post-commit gap
Projection tables and read models
Post-commit reactors and the delivery gap
Persistence, reconstruction, and crash recovery

Part V · Six providers, one product model

The ProviderAdapter contract
Drivers, instances, registries, and multi-instance routing
Codex through app-server JSON-RPC
Claude through the Agent SDK
ACP transport, Cursor, and Grok 33A. Antigravity through ACP, managed auth, and account catalogs
OpenCode ownership, recovery, and six-provider normalization
Usage accounting and provider limits without a false ledger

Part VI · The work lifecycle and integrated tools

Project discovery, onboarding import, and t3.json
Current checkout versus isolated worktrees
Start, stream, steer, interrupt, settle
Permission modes, approvals, and structured input
Threads, provider tasks, plans, skills, and subagents
Context is provider-owned; history is T3-owned
Hidden-ref checkpoints, diffs, and revert
Terminals, files, previews, MCP, VCS, and review graphs 43A. Device hosts, targets, sessions, and agent control

Part VII · Client architectures: shared semantics, platform edges

Shared runtime: connections, state, and convergence
One React renderer, three runtime edges
One thread, many deliberate projections
Electron desktop: one renderer, explicit native authority
Mobile: adaptive workspaces, persistence, and native systems

Part VIII · Reach and ship

Reachability is a route; authority is a separate proof
T3 Connect: OAuth, DPoP, relay, and tunnel
Reconnect, environments, notifications, and version skew
Distribution: artifacts, channels, and what actually ships
Release, update, and observability: three safety boundaries