Usage
Pass an optional prompt to start a session with an initial message, or launch interactively with no arguments.
Global Flags
| Flag | Short | Description |
|---|
--model <MODEL> | | Set the AI model for this session |
--permission-mode <MODE> | | Permission mode (normal, dangerous, bypass) |
--continue | -c | Resume the most recent session in the current directory |
--resume <SESSION_ID> | -r | Resume a specific session by ID |
--print [PROMPT] | -p | Print response and exit (non-interactive mode). Optionally accepts an inline prompt. |
--prompt-file <FILE> | | Load the initial prompt from a file |
--config <PATH> | | Configuration file path |
--respect-workspace-trust | | Whether to respect workspace trust settings |
devin -- add a login page
devin --model opus -- refactor the auth module
devin -c # Resume last session
devin -r abc12345 # Resume specific session
devin -p "list all TODO comments" # Print response and exit
devin -p -- list all TODO comments # Same, using -- separator (still works)
Subcommands
devin auth
Manage authentication.
| Command | Description |
|---|
devin auth login | Log in to your account |
devin auth logout | Log out and remove stored credentials |
devin auth status | Check authentication status |
devin auth login:
--force-manual-token-flow — Skip browser-based auth and manually paste a token (useful for remote/SSH sessions)
devin mcp
Manage MCP (Model Context Protocol) servers.
| Command | Description |
|---|
devin mcp add <name> | Add a new MCP server |
devin mcp list | List all configured MCP servers |
devin mcp get <name> | Show details for a specific MCP server |
devin mcp remove <name> | Remove a configured MCP server |
devin mcp login <name> | Authenticate with an MCP server via OAuth |
devin mcp logout <name> | Remove stored OAuth credentials for an MCP server |
devin mcp add:
-t, --transport <stdio|sse|http> — Transport type (optional; inferred from URL → http, trailing args → stdio)-s, --scope <local|project|user> — Configuration scope (default: local)--url <URL> — URL for HTTP/SSE transports (can also be passed as a positional argument after the name)--command <CMD> — Command for stdio transport (optional when trailing args are provided)-e, --env <KEY=VALUE> — Environment variables (repeatable)-H, --header <HEADER: VALUE> — HTTP headers (repeatable)--scopes <SCOPE,SCOPE> — OAuth scopes to request (comma-separated)<URL> — Positional URL argument for HTTP/SSE (alternative to --url)-- <COMMAND> [ARGS...] — Command and arguments for stdio (first arg is the command when --command is omitted)
Examples:
# stdio server
devin mcp add my-server -- npx @company/mcp-server --port 3000
# HTTP server (positional URL)
devin mcp add notion https://mcp.notion.com/mcp
devin mcp add --transport http datadog-mcp https://mcp.datadoghq.com/api/unstable/mcp-server/mcp
# HTTP server (--url flag, also works)
devin mcp add notion --url https://mcp.notion.com/mcp
# With environment variables and scope
devin mcp add -e GITHUB_TOKEN=ghp_xxx github -- npx -y @modelcontextprotocol/server-github
devin mcp add -s project sentry https://mcp.sentry.dev/mcp
devin mcp remove:
-s, --scope <local|project|user> — Configuration scope (default: local)
Options for devin mcp login:
--scopes <SCOPE,SCOPE> — OAuth scopes to request (comma-separated)
See MCP Configuration for details.
devin rules
Manage agent rules.
| Command | Description |
|---|
devin rules list | List all available rules |
devin rules show <name> | Show details for a specific rule |
devin rules paths | Show rule directory locations |
devin rules list:
--provider <cursor\|windsurf> — Filter by rule provider
See Rules for details.
devin skills
Manage skills.
| Command | Description |
|---|
devin skills list | List all available skills |
devin skills show <name> | Show details for a specific skill |
devin skills paths | Show skill directory locations |
devin skills list:
--trigger <user\|model> — Filter by trigger type
See Skills for details.
devin list
List sessions in the current directory. Alias: devin ls
| Command | Description |
|---|
devin list | Interactive session picker (default) |
devin list --format json | Output sessions as JSON |
devin list --format csv | Output sessions as CSV |
devin version
Print the current version and exit.
This is equivalent to devin --version.
devin acp
Run Devin as an Agent Client Protocol (ACP) server over stdio. This subcommand is intended to be invoked by an ACP-aware editor or IDE (such as Windsurf or Zed) as a subprocess — it speaks JSON-RPC over stdin/stdout and is not meant to be run interactively.
The ACP server reads credentials from WINDSURF_API_KEY if set, otherwise from the credentials stored by devin auth login. It can also accept credentials at runtime via the ACP authenticate request.
devin update
Check for updates and optionally install them.
Use --force to re-install even if already on the latest version:
devin setup
Interactive setup wizard for authentication and MCP configuration.
devin setup
devin setup --force-manual-token-flow # For remote/SSH sessions
devin uninstall
Uninstall Devin for Terminal and optionally remove all data.
| Option | Description |
|---|
--clean | Remove all data including configuration, history, and custom data |
--force | Skip confirmation prompt |
Slash Commands
These commands are available inside an interactive session. Type them at the prompt.
Mode & Model
| Command | Description |
|---|
/mode [normal|accept-edits|plan|bypass] | Show or switch the current mode |
/normal | Switch to Normal mode (default) |
/accept-edits | Switch to Accept Edits mode (auto-approve file edits in workspace) |
/plan | Switch to Plan mode (read-only planning) |
/ask <question> | Ask a question without making code changes (oneshot) |
/bypass | Switch to Bypass mode (auto-approve all actions) |
/model [name] | Show or change the current model |
/theme | Switch between themes (dark, light, 16-color, no-color) |
/bypass has aliases /yolo and /dangerous. All three do the same thing.
Session Management
| Command | Description |
|---|
/clear | Clear conversation history and start a new session. Alias: /new |
/continue [session-id] | Resume a previous session |
/fork | Fork the current session to a new session ID |
/ls [--all] | List recent sessions (current directory only by default). Aliases: /list-sessions, /resume |
/rm-session <session-id> | Irreversibly delete a session and all its data |
/exit | Exit the application (alias: /quit). You can also type exit or quit without the / prefix. |
Workspace
| Command | Description |
|---|
/workspace | List workspace directories (alias: /workspaces) |
/add-dir <path> | Add an additional workspace directory |
/undo-add-dir <path> | Remove a workspace directory |
Automation
| Command | Description |
|---|
/loop <prompt> | Run a prompt then auto-review the diff in a loop |
Extensibility
| Command | Description |
|---|
/hooks | List all loaded hooks with their IDs, event types, and source paths |
Utilities
| Command | Description |
|---|
/help | Show available slash commands |
/bug [description] | Submit a bug report |
/update [--force] | Check for and install updates. Pass --force to re-install even when already on the latest version. |
/upgrade | Upgrade your subscription plan |
/login | Authenticate with your account |
/logout | Clear stored credentials and exit |
/context | Show context window usage |
/compact | Force conversation compaction |
Modes
Modes control the agent’s autonomy level by combining a permission mode with an agent profile.
Normal (default)
Plan
Bypass
Full autonomy for complex coding tasks. The agent can read, write, and execute commands with normal permission checks.
- Permission mode: Normal
- Profile: Normal
- Use for: Multi-file refactoring, feature implementation, bug fixes
Planning only — the agent proposes changes without making them. Read-only tool access ensures no code is modified.
- Permission mode: Normal
- Profile: Plan (read-only tools)
- Use for: Architecture design, understanding codebases, planning before implementation
All permission prompts are auto-approved. The agent executes freely without asking for confirmation.
- Permission mode: Dangerous
- Profile: Normal
- Use for: Trusted tasks where interruptions slow you down
Use Bypass mode only for tasks you fully trust. All tool calls (including destructive commands) are auto-approved.
/mode, or switch directly with /normal, /accept-edits, /plan, or /bypass. Use /ask <question> as a oneshot command to ask questions without switching modes.
Profiles
Profiles determine the agent’s available tools and behavior. Profiles are automatically set when you switch modes.
| Profile | Description | Tool Access |
|---|
normal | Full coding assistant (used by Normal, Accept Edits, and Bypass modes) | All tools |
plan | Structured planning workflow (used by Plan mode) | Read-only tools (grep, glob, read, todo, ask_user_question, exit_plan_mode) |
ask | Question answering (used by the /ask command) | Read-only tools (grep, glob, read, todo, ask_user_question) |
