Anvil Desktop
anvil-app
Architecture, setup, agent sessions, work items, review, security, integrations, persistence, companion controls, and release notes.
Open docsAnvil documentation
Pick the boundary, then read the mechanics.
These docs are written for developers and maintainers. They explain what each repo owns, how to run it locally, where policy and runtime boundaries sit, and what is still alpha.
Anvil Registry
anvil-registry
Gateway architecture, local stack, package decisions, policy configuration, CLI, Admin API, LLM review, CI, deployment, and alpha limits.
Open docsAnvil Cloud
anvil-cloud
Cell contract, RuntimeRequest model, RuntimeHost boundary, local runtime, builder, Guard checks, CLI, generated client, AWS preview adapter, and v0 limits.
Open docsAnvil Node Base
anvil-registry/devcontainer-base
Safe installs, observed lifecycle-script execution, reports, strict gates, and how Node Base feeds evidence back to Anvil Registry.
Open docsFull index
Every page is markdown under content/docs.
Start here
Anvil Cloud
Architecture
Getting started
Overview
Runtime
Local runtimeHow Anvil Local runs Cells, stores local state, exposes runtime routes, and supports inspection.TestingTest Anvil Cells with in-memory adapters, mock auth, and captured logs before deploying.Builder and GuardHow the Anvil Builder checks, bundles, extracts manifests, and enforces v0 import and capability policy.Generated clientHow Anvil Cloud emits client metadata and browser helpers for calling Cell queries and mutations.
Anvil Desktop
Assurance
Engineering
ArchitectureHow Anvil Desktop separates Electron main process services, preload IPC, shared contracts, renderer UI, SQLite, and companion packages.PersistenceHow Anvil Desktop uses SQLite, schema migrations, and workspace-scoped data storage.Extend the appAdd new app capabilities without bypassing the Electron service boundaries.Build and releaseDevelopment, testing, packaging, signing, and distribution notes for Anvil Desktop.
Working guide
Operating guideSet up Anvil Desktop and run delivery workflows with real evidence.Chat and LLMConfigure AI providers, switch chat personas, and understand how Anvil Desktop grounds sessions in repository context.Agent workflowsHow to run planning, implementation, review, security, docs, BA, and handover sessions in Anvil Desktop.GitBranch management, diff review, commit preparation, and PR handover from inside Anvil Desktop.Work itemsHow Anvil Desktop brings Linear, Jira, Azure DevOps, BA notes, acceptance criteria, and implementation planning into repo-aware work.Terminal and editorUse the PTY terminal and embedded VS Code editor inside Anvil Desktop without losing workspace context.AutomationsSchedule recurring checks, run disposable worktrees, and wire automation output back into workspace evidence.Companion surfacesPair mobile, watch, Raycast, menu bar, and widget controls with the desktop app.
Anvil Registry
Concepts
ArchitectureHow Anvil Registry, the worker, persistence, object storage, queueing, CLI, Admin, and Node Base fit together.Anvil RegistryHow the npm registry gateway handles metadata, tarballs, upstreams, policy, and analysis.Policy modelHow Anvil Registry decides whether a package should be allowed, warned, quarantined, or blocked.Package decisionsInterpret allow, warn, quarantine, block, explain, analysis, LLM review, and override output.Worker analysisWhat Anvil Registry static analysis detects, how name-squatting checks work, and what evidence the worker produces.
Getting started
IntroductionWhat Anvil Registry is, who it is for, and how the pieces fit together.Alpha statusWhat the first Anvil alpha is meant for, what is working, and what still needs caution.QuickstartRun Anvil Registry locally and point npm-compatible clients at the gateway.CLIInstall and use the Anvil command-line client for package decisions, lockfile scans, reports, overrides, and operations.
Operations
Rollout guideA staged adoption path for Anvil Registry and Anvil Node Base across local trials, CI, policy tuning, and deployment.Registry configurationConfigure clients, scoped upstream registries, readiness, policy mode, and local cache behaviour.LLM integrationConfigure optional LLM risk review, understand the provider contract, and keep model output out of the enforcement seat.API referenceGateway, Admin, health, readiness, explain, policy, override, report, and package endpoint reference for alpha operators.DeploymentRun Anvil Registry locally and prepare the registry stack for AWS.Registry seedingWarm Anvil Registry from representative org lockfiles before developers and CI switch their package traffic.CI usageUse Anvil Registry and Anvil Node Base in pull requests, main branch gates, and dependency review workflows.TroubleshootingDiagnose registry routing, blocked installs, readiness failures, lifecycle scripts, and Node Base report issues.
Anvil Node Base
Anvil Node Base
Node Base safe modeInstall dependencies without lifecycle scripts and generate reviewable dependency reports.Node Base observed modeRun install scripts explicitly while capturing process, network, filesystem, and environment evidence.Node Base reportsUnderstand report files, strict-mode gates, network reports, and report submission.Network policyConfigure allowed hosts, blocked hosts, ports, and severity levels for observed mode network monitoring.