How to get my attention.

Shows the compact Discord command map, including creative, inpaint, video, and utility help.

Aliases: !command and !h. Topic forms include !help creative, !help inpaint, and !help video.

Starts a normal Alex reply using allowed memory, relationship context, recent channel history, and Alex's current voice.

Alex can also answer clear follow-ups after recent bot activity.

Lets Alex look at attached images for questions, reactions, or creative follow-ups.

Supported images can stay available briefly for follow-up questions.

Caches readable webpage, tweet, video transcript, or PDF context so Alex can summarize, answer questions, or use it in the next response.

YouTube context pairs with alex remember this video.

After a calculation returns cached Wolfram imagery, asks Alex to show a specific map, satellite view, chart, or image from that result set.

Works best immediately after alex calculate or !calculate.

Lets the room ask short continuations while recent Alex replies, attachments, links, or cached artifacts are still in context.

Keeps normal conversation moving without forcing every turn back into a fresh command.

Shows the one-screen, intent-first media quick start.

Also available as !help media, !help creative, !mediahelp, !creativehelp, and !arthelp.

Generates OpenAI images at the configured High-quality default, with compact live preview updates when the provider streams partial frames. Best for higher-fidelity stills and direct prompt-driven image work.

Supports x N (up to 3 normally or 20 for the owner), --square, --wide, --portrait, --quality=low|medium|high, and safe output-format options.

Uses xAI's current API-backed Quality model at the configured 2K default for high-fidelity batches and stylistic variation.

Supports x N (up to 3 normally or 10 for the owner), --aspect=16:9|9:16|1:1, and --resolution=1k|2k. Results retain private, expiring xAI file receipts when the provider returns them.

Generates one OpenAI High-quality image and one xAI Quality/2K image in the same response so the channel can judge the duel.

Records both artifacts with provider and model metadata.

Continues a replied or selected video using the same extension path, or your own recent video when no source is supplied.

Conversation alias for the video extension flow.

Creates a Discord-friendly sticker, cutout, or emote-style asset with crisp edges and small-size readability.

Native transparency depends on the configured OpenAI image model; Alex records when it uses a cutout fallback.

Turns Alex's last reply into an image. Works when users want Alex to visualize its own text without retyping the prompt.

Also accepts alex visualize and alex imagine that.

Edits a selected or replied-to primary image. New attachments become extra visual references; with no supplied source, Alex can use your own recent image.

The prefix form can fall back to a simple enhancement pass when no prompt is supplied. Supports safe OpenAI image options.

Runs a masked OpenAI edit using a PNG source image and matching PNG mask.

Attach source first and mask second, or attach one mask after generating an image. Alpha masks work directly; black/white masks are normalized. Use --mask-black or --mask-white when needed.

Edits one to three reference images through the pinned xAI Quality/2K route.

Good for combining references, style transfer, and provider comparison follow-ups. Supports --aspect and --resolution; completed edits retain ordered multi-parent lineage and a private Files receipt when returned.

Continues editing a selected image while preserving the artifact lineage chain. Reply to a visible artifact to fork it, or omit a source to continue from your own recent image.

Same-channel forks include creator attribution without exposing the creator's prompt or private upstream lineage.

Shows the stored prompt and edit chain for your latest image artifact or a specific owned artifact id. Foreign parent steps remain attribution-only.

Aliases: !artifact and !artifacts.

Generates xAI video from text or a deliberately supplied seed image. A standalone prompt ignores selected/recent media and defaults to 1080p on Grok Imagine Video 1.5; a replied, attached, or explicit image becomes the first frame on the same model. When that artifact carries a current private xAI Files ID, Alex reuses it before local or remote source fallback.

Text and seeded image-to-video default to 1080p and support 480p, 720p, and 1080p, preserves the seed's native aspect ratio unless --aspect is supplied, and accepts whole-second --seconds=1-15. --refs explicitly selects the 1.5 reference-to-video lane for 1-7 images and up to 15 seconds; that mode is capped at 720p.

Animates the replied or explicitly selected image. Name-addressed, real Discord mention, punctuation, and prefix forms such as alex animate that, Alex, animate that, @Alex animate that, and !animate that use the same resolver.

alex animate my last reply is intentionally different: it always uses Alex's recent text reply, even when image context exists.

Remixes a replied or selected xAI video with a new direction or visual treatment, or uses your own recent video when no source is supplied.

Visible same-channel videos can be forked with attribution.

Continues a replied or selected xAI video with new motion, framing, or scene continuation, or uses your own recent video when no source is supplied.

Alias: !continuevideo and alex continuevideo <prompt>. The prefix form can continue naturally with no prompt.

Builds Quality/2K multi-shot storyboard frames and can immediately animate a chosen shot at the seeded 1080p default. Alex renders exactly one frame per planned shot.

Supports --shots (2-4 normally or 2-6 for the owner), --animate, and --seconds. The first shot uses explicit references first; later shots use bounded first-frame and prior-frame continuity, with persisted multi-parent lineage.

The initial command plans the story and generates storyboard images with OpenAI GPT Image 2, using xAI only for technical image failures; moderation or safety blocks stop instead of provider-hopping. Review the exact narration and frames, then make the Phase 1 binary choice: Render or Cancel.

--seconds means seconds per shot. Defaults to 6 × 8-second shots; use --shots=15 --seconds=8 for 120 seconds or --shots=20 --seconds=9 for 180 seconds. Render starts xAI Orion narration and 1080p Grok Imagine Video 1.5 shots; no narration or video request starts before approval. Movie stores returned private Files receipts and ordered multi-parent lineage, retains a 1080 archive master, and delivers a bounded 720 Discord preview. Supports --narration=on|off, movie status, checkpoint-aware movie retry and delivery reconciliation, and movie retry --narration=off.

Replies to or selects one image or clip, then proposes a fixed three-pass continuation reel: three new 8-second 1080p renders, each seeded from the latest usable tail frame of the prior clip. The source media seeds the experiment but is excluded from the reel.

The Alex Movie Board shows the estimated three-request video spend and waits for approval. Recurse retains a 1080 archive master and delivers a bounded 720 preview. This experiment does not semantically rank a “best” frame, critique its own output, rewrite the direction, or rerun itself.

Animates one frame from your most recent persisted storyboard in the channel.

After a restart, Alex rebuilds recent storyboard context from stored frame artifacts. Missing shot numbers return usage help.

Shows relationship-system standings and recent movement.

Legacy aliases include !simps and !enemies.

Sentiment Market tools for individual profiles, all traders, leaderboard views, and volatility checks.

Common paths: !market me, !market @user, !market all, !market top, !market volatile. Aliases include list, ipo, leaders, and chaos.

Opens multiplayer trivia directly in #trivia with a lobby, synchronized reveals, scoring, and rematches. Games launched elsewhere use a containment thread.

Supports start, stop, leaderboard, and help.

Shows the standalone trivia scoreboard without changing the current game state.

Useful for settling the room before the next question lands.

Changes the active trivia subject mix for the current thread or next trivia run.

Pairs with !trivia help when the room wants a cleaner theme.

Reads from the server quote archive or suggests eligible-channel quotes for review.

!idiotquote accepts random, ID, or text search behavior.

Maintains the quote archive by adding eligible entries and checking the current archive size.

Useful when the server invents another quote worth preserving.

Submits a Lore Card nomination for admin review instead of pushing private server history straight onto the public record.

Use !lore help for lookup and review flows. Suggestions stay private until an admin approves visibility.

Generates fantasy-football banter in the server's own style.

Built more for server flavor than for general utility.

Runs OpenAI-backed live web search and returns a summarized result.

Role-gated in the same family as the creative tools.

Uses OpenAI Code Interpreter and can process attached files inside the sandboxed run.

Owner or approved-operator access for analysis, transforms, file work, and generated output files.

Handles calculation, unit conversion, and Wolfram-style queries through the math path.

Alex also catches the common typo form alex caculate.

Reminder intent is handled conversationally instead of as a strict prefix command.

Also recognizes set reminder, don't let me forget, wake me, and let me know when directed at Alex.

Silences Alex in the current channel for 48 hours.

Useful when a channel needs a break without changing global settings.

Fetches a GIF response through the chat path, either from an explicit search phrase or a generated contextual search.

Lives in the conversational trigger path, not the prefix registry.

Commits the current remembered YouTube or video context into long-term memory.

Also accepts alex save this video.

Reports the current long-term memory count.

Useful as a lightweight health check without opening protected tools.

Quotes a cryptocurrency by ticker symbol.

Fast market lookup for coins and tokens.

Quotes equities by ticker and serves as the main stock lookup path.

Multiple aliases exist because people type this one a few different ways.

Reserves the eR33t private server if it is idle and DMs fresh match credentials.

Built-in cooldown: five minutes between requests.

Moves Alex into or out of a voice channel.

Reserved for Team eR33t members or the owner.

Reads the Moltbook feed, semantic search results, thread structure, and available communities.

Use !molt help for the full connector map.

Runs AI feed summaries and reviews the connector's engagement audit logs.

Reserved for Team eR33t, Gentlemen, owner, or trusted equivalents.

Creates or pushes Moltbook interactions, including AI-assisted drafts and nested reply work.

Reserved for Team eR33t, Gentlemen, owner, or trusted equivalents.

Controls the Moltbook kill switch, background interaction state, dynamic interest refresh, and keyword filters.

Owner-only because these change connector behavior globally.

Triggers a stream-aware Alex reply using recent Twitch chat context, current stream metadata, and global mood.

The bot can also chime in occasionally when active chat mode is enabled.

Posts the website, Discord invite, or a lightweight global mood and status readout in Twitch chat.

Twitch-only utility commands.

Captures one stream frame and asks Alex to comment on or roast what it sees.

Creator or broadcaster only. Twitch !shot is separate from Discord !shot.

Toggles the Twitch visual watch loop for sequential stream-frame observations.

Creator or broadcaster only.