shipwithjev

Catalog / Agents & browsers

0055GitHub

TipTour

A pointer companion for the Mac, with Jev driving the cursor.

Milind Soni’s menu-bar app runs OmniParser locally on CoreML to read the screen, and Jev decides where the pointer goes. Ctrl + K types a task. It takes your own keys, and Gemini Live can drive it instead of Jev.

milind-soni/tiptour-macosREADME ↗
# TipTour

**This and That**

[](./LICENSE)
[](https://www.apple.com/macos)

</div>

A macOS menu bar companion with two modes, powered by your own API keys.

| Mode | Shortcut | What it does |
| --- | --- | --- |
| Gemini realtime | Ctrl+Option | Talk about your screen and ask for one desktop action per turn. Press again to stop. |
| JEV text | Ctrl+K | Type a click-based task. JEV chooses from locally detected controls, then TipTour executes and validates each action. |

**JEV is selected by default.** On first launch, choose JEV or Gemini, save that mode’s API key, then grant its permissions. JEV does not require microphone access. You can switch later in **Settings → Models**, which shows only the selected mode’s key field. Keys stay in macOS Keychain. There are no shared keys, hosted key proxies, or keys loaded from other projects.

Shortcuts activate only the selected mode. The primary menu-bar button and welcome hint follow your selection.

## Using TipTour

- Gemini supports clicking, typing, shortcuts, app/URL opening, scrolling, and highlighted text edits. It also supports creating an Apple Notes note from supplied content.
- JEV supports single, double, and right clicks. It reads screen labels and locations from local detection; it cannot see images or generate replacement text. Use Gemini for writing or keyboard actions.
- Press **Escape** in the text panel or click **Stop** to cancel JEV. JEV acts on its top-ranked target without a confidence cutoff. The loop stops when JEV chooses none or reports completion, on an execution pause/failure, or at its 12-action limit.
- Hold **Ctrl+Shift** and paint over an area to give Gemini focus context.
- Hold **Ctrl+Option+Command** for the Speak / Type / Highlight shortcut chooser.
- **Auto-click** lets TipTour act. **

Also filed under Agents & browsers

  1. 0295

    Flight search with Browser Use

    Breaking: Browser Use + Jev = Ultrafast ⚡ Findings flights took 7s and cost only $0.0039 🤯 > new action space every step > DOM state space > small LLM fallback to type (this video is at 1x speed btw) Built a tiny open source browser agent. try it below ↓

    @gregpr07 · Agents & browsers · ~$0.004 · ~7 s

  2. 0500

    Jev: AI Decisions as a Typed Function Call | StackToHeap

    # Jev: AI Decisions as a Typed Function Call What I learned using TypeSafe's Jev for code review, and why its more surprising uses are browser control and generative UI.…

    stacktoheap.com · Agents & browsers

  3. 0319

    x-scanner

    Chrome extension that labels every post you scroll past on X with six typed questions per post, and counts what it costs in the corner.

    oso95 · Agents & browsers

  4. 0318

    Vibe Check for X

    Chrome extension that scores draft posts and reply context before posting; optional media descriptions come from a separate vision model.

    RafalWilinski · Agents & browsers