Skip to main content

Documentation Index

Fetch the complete documentation index at: https://ade-app.dev/docs/llms.txt

Use this file to discover all available pages before exploring further.

When you open a repository in ADE for the first time, ADE creates a .ade/ directory and launches the project setup wizard. This wizard matches what you see in the app at Settings → General → Open project setup — you can re-run it any time.
If you skip setup, ADE still works. You can configure everything later through Settings. The wizard is a fast path, not a gate.

​
Step 1: Dev tools

ADE checks that your development toolchain is ready.
CheckWhat ADE looks forWhy it matters
Gitgit in PATH, version ≥ 2.20Lanes use git worktree under the hood
GitHub CLIgh in PATH and authenticatedPR management, CI status sync
Node.jsnode in PATHRequired for ADE CLI and many project workflows
If a tool is missing, ADE shows a direct link to install it. None of these checks block setup — you can proceed and install later.

​
Step 2: AI connections

Connect at least one AI provider so agents can run. ADE supports:

API key providers

Anthropic (Claude), OpenAI (GPT/Codex), OpenRouter. Paste your key and ADE verifies the connection immediately.

CLI-based providers

Claude Code CLI (claude), Codex CLI (codex), Cursor agent. ADE detects installed CLIs and checks their auth status.
ADE runs a health check on each provider. A green indicator means the provider is ready; amber means the CLI is installed but not authenticated; red means the key is invalid or the CLI is missing.
Start with one provider. You can add more any time in Settings → AI. Anthropic (Claude) is recommended for the most complete ADE experience — it works with chat, missions, CTO, and automations out of the box.

​
Step 3: Background helpers

Enable optional AI-powered background features:
FeatureWhat it doesDefault
Auto-title sessionsAI generates descriptive names for terminal and chat sessionsOn
These features use your configured AI provider and count toward your budget. Disable any you do not want.

​
Step 4: GitHub

Connect your GitHub account for PR management, CI status, and stacked PR workflows.
1

Authenticate with GitHub

ADE uses a Personal Access Token (PAT) or GitHub CLI auth. If gh is installed and authenticated, ADE detects it automatically.
2

Select repository access

Confirm which repositories and organizations ADE can access.
3

Set default branch

ADE detects your default branch (main or master) and uses it as the base for new lanes.
GitHub is optional. ADE works fully without it — you just won’t have PR management, CI sync, or stacked PR features.
ADE can build a local embedding index of your codebase for faster, more relevant context retrieval.
SettingDescription
Embedding modelLocal Xenova/all-MiniLM-L6-v2 — runs entirely on your machine, no API calls
Index scopeYour repository’s source code, docs, and configuration files
Rebuild triggersOn significant file changes, or manually from Settings → Memory
The initial indexing takes 30–120 seconds depending on repository size. During indexing, ADE remains fully usable.

​
Step 6: Linear

Connect Linear for bidirectional issue tracking between ADE and your Linear workspace.
SettingDescription
API keyLinear API key (stored in local.secret.yaml)
Sync directionBidirectional (default), read-only, or disabled
Default teamWhich Linear team receives new issues from ADE
When connected, the CTO agent can read Linear issues, update status, and create new issues. Missions can link to Linear issues for full traceability.
Linear integration is most valuable when paired with the CTO agent. The CTO can ingest new Linear issues and automatically delegate them to worker agents in dedicated lanes.

​
After setup

Once you finish (or skip) the wizard, ADE takes you to the Run tab. From here you can:

Create your first lane

Fork a task into its own isolated worktree.

Start your first agent

Use chat for a quick task or launch a mission for structured execution.

Set up managed processes

Configure your dev stack so ADE can start everything with one button.

Fine-tune settings

Adjust AI models, budgets, permissions, and notification preferences.

​
Re-running setup

You can re-run the onboarding wizard at any time:
  • Settings → General → Open project setup
  • Or navigate directly to the onboarding route in the app
Re-running setup does not reset existing configuration. It lets you review and update each step without losing your current settings.