Execution Modes
| Mode | Command | Description |
|---|---|---|
| TUI | loopy | Interactive terminal UI with chat history and spinner |
loopy -p "task" | Single task, output to stdout, exit code 0/1 | |
| REPL | loopy --repl | Simple readline loop for quick tasks |
CLI Flags
| Flag | Description | Default |
|---|---|---|
-d, --directory <path> | Working directory | Current directory |
-m, --model <name> | Model to use | Per-provider default |
-t, --max-turns <n> | Maximum agent loop iterations | 20 |
-p, --print <prompt> | Single-shot print mode | — |
--provider <name> | anthropic, openai, google-genai | anthropic |
--base-url <url> | Custom API base URL (vLLM, LM Studio) | — |
--debug | Verbose agent activity on stderr | — |
--repl | REPL mode | — |
--help | Print usage | — |
--version | Print version | — |
Slash Commands
Lines starting with/ are intercepted before reaching the agent — zero LLM tokens consumed.
| Command | Description |
|---|---|
/help | List available commands |
/clear | Clear session memory |
/quit | Exit Loopy |
/btw <question> | Stateless side question — never added to conversation history |
/skills | Discover, search, install, and manage domain skills |
/boot-new | Scaffold a new Spring Boot project from a bundled template |
/boot-setup | One-time preferences wizard (groupId, Java version, database) |
/starters | Discover Agent Starters; suggest by pom.xml triggers |
/boot-add | Bootstrap domain capabilities into an existing project |
/boot-modify | Structural modifications (Java version, native support, CI) |
/forge-agent --brief <path> | Bootstrap an agent experiment project from YAML brief |
/session save [name] | Save current session |
/session list | List saved sessions |
/session load <id> | Restore a saved session |
Agent Tools
The embedded agent has access to these tools:| Tool | Description |
|---|---|
Bash | Execute shell commands |
Read | Read file contents |
Write | Create or overwrite files |
Edit | Targeted edits to existing files |
Glob | Find files by pattern |
Grep | Search file contents |
ListDirectory | List directory with optional depth |
Skill | Load domain skills on demand |
Submit | Submit final answer (ends the loop) |
TodoWrite | Track work items |
Task | Delegate to subagents |
TaskOutput | Retrieve background subagent results |
AskUserQuestion | Ask user for clarification (TUI only) |
WebSearch | Web search (requires BRAVE_API_KEY) |
WebFetch | Fetch and summarize web pages |
Environment Variables
| Variable | Description |
|---|---|
ANTHROPIC_API_KEY | Anthropic API key (required for default provider) |
OPENAI_API_KEY | OpenAI API key |
GOOGLE_API_KEY | Google Gemini API key |
BRAVE_API_KEY | Brave Search API key (for WebSearch/WebFetch) |
LOOPY_DEBUG_LOG | Path for debug log file |
Default Models
| Provider | Model |
|---|---|
| Anthropic | claude-sonnet-4-20250514 |
| OpenAI | gpt-4o |
| Gemini | gemini-2.5-flash |
--model or .model() in the programmatic API.