# AI use cases for Agent Hud Native Open (jazzenchen/agent-hud-native-open)

Repo: [jazzenchen/agent-hud-native-open](https://github.com/jazzenchen/agent-hud-native-open) | Use cases: 4 | Page: https://dedupedai.com/use-cases/jazzenchen-agent-hud-native-open/

Agent HUD's native Swift app for macOS, which shows usage, quotas, tokens and sessions for Claude Code, Codex and 16 more coding agents at the Mac's notch or a screen edge.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Watch agent quotas from the screen edge while you work: https://dedupedai.com/use-cases/jazzenchen-agent-hud-native-open/glance-at-agent-quotas-while-working/
- Answer agent approvals and questions from the HUD: https://dedupedai.com/use-cases/jazzenchen-agent-hud-native-open/answer-agent-approvals-without-switching-windows/
- Review 30 days of token spend by agent and model: https://dedupedai.com/use-cases/jazzenchen-agent-hud-native-open/review-token-spend-by-agent-and-model/
- Preview the HUD with sample data before adding agents: https://dedupedai.com/use-cases/jazzenchen-agent-hud-native-open/try-the-hud-with-sample-data/

## How to watch agent quotas from the screen edge while you work with Agent Hud Native Open

Repo: [jazzenchen/agent-hud-native-open](https://github.com/jazzenchen/agent-hud-native-open) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README

Keep each quota's green, yellow or red glow in view and hover for the reset countdown.

**Who it is for:** A developer running Claude Code or Codex all day who keeps switching apps and needs to know how much quota is left without opening each agent.

**The problem:** Quota limits are easy to hit unexpectedly, and checking each agent's usage page breaks focus.

**The outcome:** A glow at the notch or on a screen edge shows each quota's level, and a hover shows the share used and reset countdown.

### Steps

1. Clone the repository with `git clone https://github.com/jazzenchen/agent-hud-native-open.git`
2. Enter the folder with `cd agent-hud-native-open`
3. Run `make check` and then `make test` to confirm the build and tests pass
4. Run `make run` to build and open the app
5. In Settings, choose the accounts and quotas to show, then hover the glow to read the numbers

### Prompt to paste into your agent

```text
Clone jazzenchen/agent-hud-native-open, run make check, make test and make run on the Mac, then report whether build/Agent HUD Open.app opened and which quotas appear for <agent name>. Verify that the report lists a share used and a reset time for at least one quota.
```

### Caveat

Quotas for Copilot need your consent in Settings and Codex approvals need its hook trusted in Codex, so some readings stay blank until you set those up.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/jazzenchen-agent-hud-native-open/glance-at-agent-quotas-while-working/

## How to answer agent approvals and questions from the HUD with Agent Hud Native Open

Repo: [jazzenchen/agent-hud-native-open](https://github.com/jazzenchen/agent-hud-native-open) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Allow, deny or answer an agent's request on the HUD and jump back to its session with one click.

**Who it is for:** A developer running several agents in parallel who gets interrupted by approval prompts buried in other windows or terminal tabs.

**The problem:** An agent waiting for an approval or a question stalls until someone finds the right window.

**The outcome:** Pending requests appear on the HUD with the command, file or diff, and can be answered in place.

### Steps

1. Complete the build steps: `git clone https://github.com/jazzenchen/agent-hud-native-open.git`, then `make run`
2. Confirm the agent's hooks were written, which Agent HUD does on your Mac
3. When an agent asks, read the command, file or diff on the HUD
4. Choose Deny, Allow once or Always allow, or pick an option for a question
5. If you need the context, click the session title or Return to session

### Prompt to paste into your agent

```text
In the agent-hud-native-open repository, start the app with make run and trigger an approval in <agent name>. Report the request text shown on the HUD and the buttons offered. Verify that answering on the HUD lets the agent continue.
```

### Caveat

Agent HUD does not answer for you, so a request you ignore returns to the agent's own prompt after 10 minutes, and Always allow applies the agent's rule going forward.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/jazzenchen-agent-hud-native-open/answer-agent-approvals-without-switching-windows/

## How to review 30 days of token spend by agent and model with Agent Hud Native Open

Repo: [jazzenchen/agent-hud-native-open](https://github.com/jazzenchen/agent-hud-native-open) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Open the statistics window to see tokens and API-priced cost by time, agent, model and session.

**Who it is for:** A developer or team lead who wants to know which agent, model or session is using the most tokens and what it would cost at API prices.

**The problem:** Token use is spread across several agents and sessions, so the total cost and the heavy sessions are hard to see.

**The outcome:** A statistics view showing token history by time, agent, model and session, with an estimated cost.

### Steps

1. Build and open the app with `make run`
2. Let your agents run so the app collects token history, which is kept for 30 days on your Mac
3. Open the statistics window and choose a time range of 5 hours, 24 hours, 7 days or 30 days
4. Switch to the by agent and model view and narrow the chart to the models and token kinds you need
5. Open a session to see tokens per turn and how full the context window got

### Prompt to paste into your agent

```text
Use agent-hud-native-open on this Mac after make run to open the statistics window and read the 30 day token totals for each agent. Report the top three models by tokens and their estimated cost. Verify that costs marked with the approximate symbol are noted as estimates.
```

### Caveat

Costs are priced at API prices where a model has one and are only an estimate on a subscription, not a charge.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/jazzenchen-agent-hud-native-open/review-token-spend-by-agent-and-model/

## How to preview the HUD with sample data before adding agents with Agent Hud Native Open

Repo: [jazzenchen/agent-hud-native-open](https://github.com/jazzenchen/agent-hud-native-open) | Difficulty: Beginner | Time: 15 minutes | Basis: From the README

Open the app on sample data to see every mode and panel before you sign in to any agent.

**Who it is for:** A developer evaluating Agent HUD who wants to see the layout, modes and statistics before connecting real accounts.

**The problem:** Setting up agents first costs time, and it is unclear what the HUD looks like or whether it fits the workflow.

**The outcome:** The app opens on sample data, showing the glow, hover panel, statistics and approval layout without real accounts.

### Steps

1. Clone the repository with `git clone https://github.com/jazzenchen/agent-hud-native-open.git`
2. Enter the folder with `cd agent-hud-native-open`
3. Run `make build` to build the locally signed app
4. Quit the app if it is open, then run `make demo`
5. Switch between the Notch, Simple and Dynamic modes in Settings to compare them

### Prompt to paste into your agent

```text
In the agent-hud-native-open repository, run make demo and report the three display modes available in Settings. Verify that the app shows sample data and that no real agent account is read.
```

### Caveat

The demo needs the Xcode toolchain to build, so a non-developer will need someone to set that up before a copy and paste run works.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/jazzenchen-agent-hud-native-open/try-the-hud-with-sample-data/
