WaitPro by Maximem

Turn AI wait time into reading time.

Every time you send a prompt to Claude Code, you wait. WaitPro fills that gap with a short, useful read in your sidebar, then gets out of the way the moment the reply is ready.

• Ready before you ask: your next reads are claimed ahead of time, so the card appears the moment you send a prompt, with no loading step
• Opens the instant you send a prompt, closes the instant the turn ends
• Answer-first reads: the headline point up top, then the key points, then the detail, with a "Read more" link
• Long wait? Next ▸ moves on to another read (and counts the current one as read); Skip ✕ is for reads you don't want: it is skipped for good
• Never repeats: each read is marked as read when the turn finishes
• Never steals focus: your typing stays where it was
• One card, not two: while the extension is running, the terminal status-line card is hidden in your IDE's terminal (the sidebar shows it instead)
• Your call: close it mid-turn and it stays closed until your next prompt

What you'll read

WaitPro picks something for you in this order:

1. Your own bookmarks. Chrome and X bookmarks you've connected to Maximem, turned into short summaries.
2. Topics you follow. No bookmarks yet, or you've read them all? WaitPro shows a recent article on a topic you care about, chosen from what Maximem already knows about you (your top few topics). You don't need any bookmarks for this.
3. Popular topics. New to Maximem, with no bookmarks and no memory yet? WaitPro starts you on a few popular topics (AI and machine learning, startups, software engineering, productivity, science and technology) until it learns yours. Those reads say "Popular topic", never "you follow". Once Maximem has topics of your own, WaitPro switches to them, unless you've already changed your list.
4. Nothing yet. If a topic's articles are all read, the card says "Nothing new to read" and fills back in as new articles arrive.

You send a prompt ──▶ WaitPro asks Maximem for a read
                        ├─ unread bookmark?          → show it
                        ├─ else a topic you follow?  → show a recent article
                        ├─ else a popular topic?     → show a recent article
                        └─ else                      → "Nothing new to read"
Reply finishes ────────▶ card closes, read is marked as read

Get started

1. Install the extension.
2. Open the WaitPro sidebar and click Sign in with API key (or run WaitPro: Sign in from the Command Palette). Paste a key from https://app.maximem.ai.
3. Restart any open Claude Code session so it loads the WaitPro hooks.
4. Send a prompt. The card appears while Claude works.

New to Maximem? Create an account at app.maximem.ai first. To get personal reads, either connect your Chrome or X bookmarks in the Maximem dashboard and turn WaitPro on in Account Settings, or just use Maximem with your AI chats so it learns your topics.

Requires Claude Code, on macOS or Linux. WaitPro reacts to Claude Code only: other agents (for example the built-in Gemini agent in Antigravity, or Copilot chat) don't send it any signal, so the card won't open for them. WaitPro: Sign out removes your key.

Tips

• Card appears in a side tab instead of the sidebar? Some editors keep the WaitPro view hidden. WaitPro then shows the same card in a side editor tab and closes it when the reply is ready. Set waitpro.surface to panel to always use the tab.
• Keep your Explorer free. WaitPro lives in the left sidebar by default. Right-click its icon, choose Move to Secondary Side Bar, and set waitpro.location to secondary.
• Seeing "Nothing new to read"? You've read everything currently available. It fills back in as new articles arrive; adding bookmarks in the Maximem dashboard gives WaitPro more of your own reads.

Settings

Privacy

• Reads are fetched by the WaitPro hooks for Claude Code, which talk to the Maximem API with your key. The extension only shows what the hooks write to disk.
• Anonymous usage analytics. The extension and the hooks send event names (activated, card shown, served, next, skipped, read) with a few flags (surface, app version, OS, whether a read was curated, how long a card was on screen) to Maximem's telemetry API. Never your key in the body, your email, or the text, title or URL of a read. It follows your editor's telemetry setting (off means nothing is sent from the extension), and WAITPRO_ANALYTICS=0 in ~/.claude/waitpro/config/config.env turns it off for the hooks too.
• Your memory picks topics, nothing more. For topic reads, Maximem uses your memory only to choose topic names. Article summaries are made without any of your personal context.
• It writes to your machine only when you sign in: your key to ~/.claude/waitpro/credentials (readable only by you), the hook scripts under ~/.claude/waitpro/, and the WaitPro entries in ~/.claude/settings.json (backed up first; your other settings and hooks are left alone). Sign out removes the key.
• Your key is never stored in VS Code settings and never logged.
• It also refreshes a small heartbeat file (~/.claude/waitpro/ui.json) every few seconds so the terminal card knows to stay hidden. If the IDE closes, the terminal card comes back.
• To have your next read ready, the hooks keep a per-session buffer of up to 10 upcoming reads in ~/.claude/waitpro/prefetch/ (removed when the session ends). Reads you never see simply return to your queue after about 30 minutes. WAITPRO_PREFETCH_DEPTH changes how many.
• Otherwise it only reads the small state files the hooks write under ~/.claude/waitpro/sessions/, and shows the text in a sandboxed panel (no remote content, text escaped, http(s) links only).

Links

• Website: https://maximem.ai
• Get an API key: https://app.maximem.ai