.{ // MAX_HELP_LEN| .general = \\usage: {0s} [arguments] \\ \\The commands are: \\ agent starts an interactive AI agent that can browse the web \\ fetch fetches the specified URL \\ help displays this message \\ mcp starts an MCP (Model Context Protocol) server (stdio or HTTP) \\ run runs a saved script (no LLM), then exits \\ serve starts a WebSocket CDP server \\ version displays the version of {0s} \\ \\Use "{0s} help " for more information about a command. , .serve = \\usage: {0s} serve [OPTIONS] [COMMON_OPTIONS] \\ \\Starts a WebSocket CDP server. \\ \\options: \\ --advertise-host \\ The host to advertise, e.g. in the /json/version response. Useful, \\ for example, when --host is 0.0.0.0. \\ Defaults to --host value. \\ --cdp-max-connections \\ Maximum number of simultaneous CDP connections. \\ Defaults to 16. \\ --cdp-max-http-message-size \\ Maximum allowed HTTP request size \\ Defaults to 4096 (maximum allowed: 16383) \\ --cdp-max-message-size \\ Maximum allowed incoming websocket message size. \\ Defaults to 1048576 (1MB) \\ --cdp-max-pending-connections \\ Maximum pending connections in the accept queue. \\ Defaults to 128. \\ --cookie \\ Path to a JSON file to load cookies from (read-only). \\ Defaults to no cookie loading. \\ --disable-metrics \\ Disables the /metrics endpoint (Prometheus text format). \\ Defaults to false (the endpoint is exposed). \\ --host \\ Host of the CDP server. \\ Defaults to "127.0.0.1". \\ --port \\ Port of the CDP server. \\ Defaults to 9222. , .fetch = \\usage: {0s} fetch ... [OPTIONS] [COMMON_OPTIONS] \\ \\Fetches one or more URLs. Each URL is loaded in its own page. Passing more \\than one URL requires --json. \\ \\options: \\ --cookie \\ Path to a JSON file to load cookies from (read-only). \\ Defaults to no cookie loading. \\ --cookie-jar \\ Path to a JSON file to save cookies to on exit (write-only). \\ Defaults to no cookie saving. \\ --dump \\ Dumps the document to stdout. \\ Defaults to no dump. \\ Allowed values: \\ html Serialized HTML of the DOM. \\ markdown Converts content to Markdown. \\ png Text-only rendering of the page as a \\ PNG image (base64 with --json). \\ semantic_tree JSON-serialized semantic tree. \\ semantic_tree_text Pruned plain-text semantic tree. \\ --inject-script \\ JavaScript to execute as the document's is parsed, before any \\ other scripts in the page run. Can be passed multiple times; scripts \\ run in order. \\ --inject-script-file \\ Like --inject-script, but reads the script from a file. Can be passed \\ multiple times; can be mixed with --inject-script and runs in CLI order. \\ --json \\ Capture and print the status of the fetch as JSON. A single URL \\ prints one object; multiple URLs print {{"results": [ ... ]}} with \\ one object per URL. When used with --dump the dumped \\ content is wrapped within each object. \\ --dump-max-bytes \\ Soft cap on the dumped html or markdown, in bytes. Output is cut \\ at a UTF-8 boundary and a '[truncated]' marker is appended. \\ Defaults to no cap. \\ --metrics \\ Write metrics (Prometheus text format) to stdout on exit, after \\ any --dump output. \\ Defaults to false. \\ --dump-duselector \\ Dump only the first element matching the CSS selector instead \\ of the whole document. Fails if nothing matches. \\ --strip-mode \\ Tag group to remove from dump. Can be passed multiple times. \\ In markdown only 'ui' changes the output: scripts, styles and \\ hidden elements are never rendered. \\ Defaults to no-strip. \\ Allowed values: \\ js Script and link[as=script, rel=preload]. \\ ui Includes img, picture, video, CSS and SVG. \\ css Includes style and link[rel=stylesheet]. \\ invisible Best-effort (e.g. display:none) hidden elements \\ full Strip everything. \\ --terminate-ms \\ Hard deadline in milliseconds. After this time elapses, JavaScript \\ execution is forcibly terminated (e.g. for pages with endless scripts). \\ Unlike --wait-ms, which only stops waiting, --terminate-ms aborts the \\ page. \\ Defaults to no terminate. \\ --wait-ms \\ Wait time in milliseconds. Supersedes all other --wait parameters. \\ Defaults to 5000. \\ --wait-script \\ Wait for a JavaScript expression to return truthy. Checked after \\ --wait-until condition is met. \\ --wait-script-file \\ Like --wait-script, but reads the script from a file. \\ --wait-selector \\ Wait for an element matching the CSS selector to appear. Checked after \\ --wait-until condition is met. \\ --wait-until \\ Wait until the specified event. Checked before other --wait-* options. \\ Defaults to 'load'. If --wait-selector, --wait-script or \\ --wait-script-file specified, defaults to none. \\ Allowed values: "load", "domcontentloaded", "networkalmostidle", \\ "networkidle", "done". \\ 'done' waits for full quiescence (no pending scripts or network); \\ pages with constant background activity never reach it and run \\ to --wait-ms. \\ --with-base \\ Add a tag in dump. \\ Defaults to false. \\ --with-frames \\ Includes the contents of iframes. \\ Defaults to false. , .mcp = \\usage: {0s} mcp [OPTIONS] [COMMON_OPTIONS] \\ \\Starts an MCP (Model Context Protocol) server. Defaults to stdio; pass \\--port to serve over HTTP with an independent browsing session per client. \\ \\options: \\ --cdp-port \\ Also run a CDP (WebSocket) server on this port. Cannot be \\ combined with --port (they share one network listener). \\ Defaults to disabled. \\ --cookie \\ Path to a JSON file to load cookies from (read-only). \\ Defaults to no cookie loading. \\ --cookie-jar \\ Path to a JSON file to save cookies to on exit (write-only). \\ Defaults to no cookie saving. \\ --host \\ Host to bind when --port is set (e.g. 0.0.0.0 for all interfaces). \\ Defaults to "127.0.0.1". \\ --port \\ Serve MCP over HTTP on this port instead of stdio. Clients POST \\ JSON-RPC to /mcp; route requests to a session with the \\ Mcp-Session-Id header. \\ Defaults to stdio (no HTTP server). , .agent = \\agent command \\Starts an interactive AI agent that can browse the web. \\ \\Usage: \\ {0s} agent [SCRIPT] [OPTIONS] [COMMON_OPTIONS] \\ \\Examples: \\ {0s} agent (auto-detects API key from env) \\ {0s} agent --provider anthropic --model claude-sonnet-4-6 \\ {0s} agent --provider huggingface (HF serverless router, HF_TOKEN) \\ {0s} agent --provider ollama --model qwen3.5:latest \\ {0s} agent --no-llm (basic slash-command-only REPL) \\ {0s} run script.js (replay a saved script; see `run`) \\ {0s} agent --task "..." --save out.js (synthesize a replayable script) \\ \\Arguments: \\[SCRIPT] \\ Optional path to a .js script. Runs the script (no LLM calls) and \\ exits; `{0s} run SCRIPT` is the preferred spelling. With no script \\ and no --task, the REPL starts; from there /load runs a script and \\ /save exports the session to a file. \\ Caution: .js files can contain evaluate(...) calls that run \\ arbitrary JavaScript in the page. Only run scripts you trust, the \\ same way you would a shell script. \\ \\Options: \\ -a, --attach \\ Feed a local file to the model alongside --task. Repeatable, one \\ file per flag. Text files are inlined (max 512 KiB each); \\ images/audio/pdf are base64-encoded (max 20 MiB each). \\ Requires --task. \\ --base-url \\ Override the API base URL for the provider. Defaults to the \\ provider's standard endpoint. \\ Ollama default: http://localhost:11434/v1. \\ llama.cpp default: http://localhost:8080/v1. \\ Hugging Face default is the serverless router \\ (https://router.huggingface.co/v1); \\ point this at a dedicated Inference Endpoint to use one. \\ --effort \\ Per-turn reasoning budget, mapped to each provider's native \\ thinking/reasoning knob. Default: low in the REPL (snappy turns), \\ medium in one-shot --task mode, unless the provider sets its own \\ default (Mistral defaults to none, as its default model rejects \\ effort). In the REPL, use /effort to change it. \\ Allowed values: none, minimal, low, medium, high, xhigh. \\ --list-models \\ Print the model IDs usable with `agent` for --provider, one per \\ line, sorted, and exit. Auto-detects the provider from env when \\ --provider is omitted. \\ --model \\ The model name to use. Defaults to a sensible default per \\ provider. In the REPL, use /model to list and change models for \\ the active provider. \\ --no-llm \\ Force the basic REPL even when an API key is present or \\ --provider is set. Useful for testing slash commands without \\ burning tokens, or for disabling the LLM in a saved command \\ without editing the existing flags. Wins over --provider. \\ --provider \\ The AI provider. When omitted, lightpanda auto-detects an API \\ key from your environment (ANTHROPIC_API_KEY, OPENAI_API_KEY, \\ GOOGLE_API_KEY/GEMINI_API_KEY, HF_TOKEN, AI_GATEWAY_API_KEY, \\ MISTRAL_API_KEY). With exactly one key set: that provider is \\ used. With multiple keys on a TTY: you'll be prompted to pick; \\ in non-interactive contexts, pass --provider explicitly. With \\ no keys set: falls back to the basic REPL (slash commands only, \\ no natural-language input, no LOGIN / ACCEPT_COOKIES keywords). \\ \\ Local servers (ollama, llama_cpp) are never auto-detected (they \\ need no key); select them explicitly with --provider ollama / \\ --provider llama_cpp. \\ \\ Allowed values: "anthropic", "openai", "gemini", "huggingface", \\ "vercel", "mistral", "ollama", "llama_cpp". \\ In the REPL, use /provider to list and change providers. \\ --save \\ Synthesize a replayable .js script from the --task run and write \\ it to PATH, instead of printing the answer. Replay it later with \\ `run PATH` (no LLM calls). Overwrites PATH if it exists. \\ Requires --task. \\ --system-prompt \\ Override the default system prompt. \\ --task \\ One-shot mode: run a single user turn, print the final answer \\ to stdout, and exit. Conflicts with the positional script. With \\ --save, the answer is suppressed and a script is written instead. \\ --verbosity \\ Stderr chatter level. Default: high when --task captures stderr \\ to a pipe or file; low otherwise. low/medium also raise \\ --log-level to err (mutes page-side console.error spam) unless \\ --log-level is set explicitly. \\ Allowed values: \\ low silent in --task mode (final answer to stdout only); \\ spinner + summary in REPL. \\ medium + one `● [tool: ...]` line per call. \\ high + the matching `[result: ...]` body (required by the \\ benchmarks harness). \\ \\The provider, model, effort, and verbosity you choose in the REPL are \\remembered per-directory in .lp-agent.zon and reused on the next run. \\ \\API keys are read from the environment: ANTHROPIC_API_KEY, OPENAI_API_KEY, \\GOOGLE_API_KEY/GEMINI_API_KEY, HF_TOKEN, AI_GATEWAY_API_KEY, or \\MISTRAL_API_KEY. The local servers (Ollama, llama.cpp) do not require an \\API key. , .run = \\run command \\Runs a saved script, then exits. No LLM calls, no API key needed. \\ \\Usage: \\ {0s} run