# AI use cases for Bough (nickelsec/bough)

Repo: [nickelsec/bough](https://github.com/nickelsec/bough) | Use cases: 4 | Page: https://dedupedai.com/use-cases/nickelsec-bough/

bough reads your Claude Code, Codex CLI and Pi session transcripts locally and shows what you built, broken into days, tasks and prompts, with what each piece would have cost at published API rates.

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

- Price each feature built with an AI coding agent: https://dedupedai.com/use-cases/nickelsec-bough/price-each-feature-you-built-with-an-agent/
- Get a plain text cost summary for a project: https://dedupedai.com/use-cases/nickelsec-bough/get-a-plain-text-cost-summary-for-a-project/
- Export agent work history as JSON for other tools: https://dedupedai.com/use-cases/nickelsec-bough/export-agent-work-graph-as-json/
- Compare agent costs between Claude Code and Codex: https://dedupedai.com/use-cases/nickelsec-bough/compare-costs-across-claude-code-and-codex/

## How to price each feature built with an AI coding agent with Bough

Repo: [nickelsec/bough](https://github.com/nickelsec/bough) | Difficulty: Beginner | Time: 15 minutes | Basis: From the README

See the rough API-rate cost of each task inside a project, not just each session.

**Who it is for:** A developer who uses Claude Code, Codex or Pi daily and wants to know which features or bugs cost the most when the agent did the work.

**The problem:** Usage tools only report cost per session, and one session can cover weeks of unrelated work, so the spend cannot be tied to things you recognise.

**The outcome:** A browser view of your projects with time at the keyboard and cost for each, where you can open a project and see the cost of each task and each prompt.

### Steps

1. Install Bough with `brew install nickelsec/tap/bough` on macOS, or use the install script for your platform from the README.
2. Run `bough` to open your projects in the browser.
3. Open a project and hover a task to see its token and dollar breakdown.
4. Click a task to open the full prompt record beside the drawing.

### Prompt to paste into your agent

```text
Using the bough repo at github.com/nickelsec/bough, run `bough <project-name> --text` for the project at <project-name> and list the three most expensive tasks with their dollar figures. Verify that each figure is labelled as API-rate cost, not a bill.
```

### Caveat

Costs are priced at published API rates for the model used, so a flat-rate subscription will not match these figures.

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/nickelsec-bough/price-each-feature-you-built-with-an-agent/

## How to get a plain text cost summary for a project with Bough

Repo: [nickelsec/bough](https://github.com/nickelsec/bough) | Difficulty: Beginner | Time: 15 minutes | Basis: From the README

Print a project's sittings, tasks and costs to the terminal or a file.

**Who it is for:** A developer or team lead who wants a quick written record of how much agent work went into a project, without opening a browser.

**The problem:** The browser view is hard to paste into notes or a status update, and a plain list is quicker to scan.

**The outcome:** An indented text list with prompt counts, changes to files, hours at the keyboard, commits and the API-rate total, saved to a file if you want it.

### Steps

1. Install Bough as described in the README.
2. Run `bough --list` to see every project with history.
3. Run `bough <project-name> --text` to print the summary to the terminal. (inferred, check the README)
4. Run `bough <project-name> --text -o notes.txt` to save it to a file. (inferred, check the README)

### Prompt to paste into your agent

```text
In the bough repo at github.com/nickelsec/bough, run `bough --list` and then `bough <project-name> --text -o summary.txt` for the project named <project-name>. Confirm that summary.txt exists and contains a total dollar figure labelled as API-rate cost.
```

### Caveat

The text view counts what the agent did from its transcripts, so it may not match the number your agent host shows, and Bough says so in its output.

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/nickelsec-bough/get-a-plain-text-cost-summary-for-a-project/

## How to export agent work history as JSON for other tools with Bough

Repo: [nickelsec/bough](https://github.com/nickelsec/bough) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Write the full task and cost structure of a project to JSON for your own analysis.

**Who it is for:** A developer or analyst who wants to feed agent session costs into a spreadsheet, script or dashboard.

**The problem:** The browser and text views cannot be processed by other programs, so the structure of the work stays locked in the tool.

**The outcome:** A JSON file holding the sittings, tasks and prompts for a project, with facts about the work and no layout data.

### Steps

1. Install Bough as described in the README.
2. Run `bough <project-name> --json -o work.json` to write the graph to a file. (inferred, check the README)
3. Open work.json in a script or spreadsheet and total the cost fields per task.
4. Run `bough <project-name> --json --no-repo -o work.json` if you do not want Bough to read the project's git history. (inferred, check the README)

### Prompt to paste into your agent

```text
Using the bough repo at github.com/nickelsec/bough, run `bough <project-name> --json -o work.json` for the project <project-name>. Check that work.json parses as valid JSON and that it contains task entries with cost fields.
```

### Caveat

The JSON carries no colours, sizes or positions, only what Bough judges true about the work, so a consumer must do its own layout.

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/nickelsec-bough/export-agent-work-graph-as-json/

## How to compare agent costs between Claude Code and Codex with Bough

Repo: [nickelsec/bough](https://github.com/nickelsec/bough) | Difficulty: Beginner | Time: 15 minutes | Basis: From the README

Filter a project's history to one agent at a time to compare what each cost.

**Who it is for:** A developer who uses both Claude Code and Codex CLI and wants to see where each one's work went and what it cost.

**The problem:** Mixed history from two agents in one project list makes it hard to judge which agent handled which work.

**The outcome:** Separate views of the same projects for Claude Code and for Codex, each with its own cost totals.

### Steps

1. Install Bough as described in the README.
2. Run `bough --agent=claude` to show only Claude Code work.
3. Run `bough --agent=codex` to show only Codex CLI work.
4. Compare the totals for the same project in each view and note any figures that look wrong, as the README asks.

### Prompt to paste into your agent

```text
In the bough repo at github.com/nickelsec/bough, run `bough --agent=claude --text` and `bough --agent=codex --text`. Report the total API-rate cost from each output and state which agent's figures the README says deserve more suspicion.
```

### Caveat

The README says Codex and Pi support is newer and has had less exposure than Claude Code, so Codex figures deserve more checking.

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/nickelsec-bough/compare-costs-across-claude-code-and-codex/
