A local-first desktop app that generates realistic, fake customer-support tickets with an LLM — driven by a handful of numeric settings and a fully editable prompt. It's useful for seeding demos, load-testing a helpdesk UI, or producing sample data without touching real customer information.
Everything runs on your machine. The app only reaches out to the network to make the LLM call itself; generated tickets are written to a plain JSON file you can view in-app and export.
Highlights
- Two providers: Ollama (local, default) and Anthropic. (OpenAI and Google Gemini support is planned for a future release.)
- Hosted API keys are stored in your OS keychain (via Electron
safeStorage) — never in plaintext config, never exposed to the UI. - A monochrome, neo-brutalist interface with a paginated ticket viewer and per-ticket conversation threads.
- Output is a self-describing
tickets.json(tickets + run metadata: token usage, cost, timing).
- Go to the Releases page and download the latest
.dmg. - Open the
.dmgand drag Qbort into your Applications folder. - First launch (unsigned app): the app is currently unsigned, so macOS blocks it the first time you open it ("Apple cannot check it for malicious software"). The exact steps depend on your macOS version. On macOS 14 and earlier, right-click (or Control-click) the app → Open, then click Open in the confirmation dialog. On macOS 15 (Sequoia) and later, double-click the app once (it will be blocked), then open System Settings → Privacy & Security, find the "Qbort was blocked" message near the bottom, and click Open Anyway. You only need to do this once. Either way, this Terminal command clears the quarantine flag and skips the dialogs entirely:
xattr -dr com.apple.quarantine "/Applications/Qbort.app"
Requires Node.js 20+ and npm.
git clone https://github.com/balevine/qbort.git
cd qbort
npm install
# Run the app in development (hot reload):
npm run dev
# — or — build a distributable (.dmg + .zip land in release/):
npm run dist:macHandy scripts: npm test (unit + integration suite), npm run typecheck.
The app is a single screen: a ticket viewer in the middle, and a top bar with LOAD TICKETS, GENERATE, and a gear (settings) button. All configuration lives behind the gear.
Open Settings → Provider. The default provider is Ollama so the app works out of the box with no API key.
- Make sure your Ollama server is running (default host
http://localhost:11434). - In the Provider panel, confirm the host and click Fetch models, then pick an installed model from the dropdown.
That's it — local generation is free and never leaves your machine. (Managing/installing Ollama models themselves is outside the scope of this app.)
- Select Anthropic as the provider.
- Paste your API key and click Save. The key is encrypted into your OS keychain; the app only ever shows whether a key is set, never the value. Use Test connection to verify it.
- You do not pick a model. For hosted providers the app automatically uses a curated, cost-balanced mid-tier model — for Anthropic, this is Claude Sonnet 5, used by default and not user-changeable. Synthetic ticket generation is high-volume and doesn't need a flagship reasoning model, so the choice is fixed to keep runs fast and cheap.
Open Settings (gear icon). The generation settings are:
| Setting | Default | Range | What it does |
|---|---|---|---|
| Number of tickets | 100 | 1 – 5000 | How many tickets to generate in a run. |
| Include staff responses | off | — | When off, each ticket has only the customer's opening message. |
| Average staff responses | 0 | 0 – 20 | Mean replies per ticket; the actual count varies around this. |
| Number of staff members | 10 | 1 – 100 | Size of the staff roster used to author replies. |
| Max ticket age (days) | 90 | 1 – 3650 | How far back message timestamps are spread — each ticket opens somewhere in this window, with replies following later. |
Staff roster. A staff member is a name + alias; their email is derived as alias@company.biz. The app ships with a default roster and resizes it as you change the staff count. This roster is the source of truth for who authors staff replies, and the @company.biz domain is how the viewer tells staff apart from customers.
Default folder. Under Settings → Storage you can pick the folder where ticket files are saved and auto-loaded from. If unset, it defaults to the app's user-data directory.
There is no save button — settings persist automatically as you change them (written to a settings.json in the app's user-data directory). Just close the modal when you're done.
This is where the generator is most customizable. Under Settings → Prompt you'll find a large editor holding the creative half of the prompt: ticket categories and their percentages, sentiment mix, tone, product/domain context — whatever shapes the tickets you want. It ships with a minimal starter prompt that you expand on based on what you need.
At generation time the app compiles the final prompt: your text, followed by a clearly delimited block of app-enforced requirements (the exact JSON schema, how many tickets this batch should produce, the allowed status values, and the staff-response rules). Your text is never silently overridden — the requirements are appended. Use Preview compiled prompt to see exactly what gets sent to the model.
- Click GENERATE in the top bar.
- You'll see a cost estimate first — estimated tokens and cost (Ollama shows
$0 · local). Nothing runs until you confirm. - Confirm to start. Progress streams live (tickets done, batches, retries) and you can Cancel at any time — tickets produced so far are still saved.
- When the run finishes, the tickets are written to
tickets.jsonin your default folder and loaded into the viewer. The file also records the run's real token usage, cost, and duration.
Click LOAD TICKETS in the top bar to open any tickets JSON file. On launch, the app also auto-loads the most recent file from your default folder, so your last run is there waiting.
The viewer is a paginated table (100 per page) with columns for id, subject, status, from, and created (the ticket's opening time). You can:
- Filter by status and search free-text across the subject and every message's body and author.
- Click any row to open the conversation modal — the full thread rendered as messages (each with its author and timestamp), customer messages on white and staff messages on slate-grey.
- Read the summary at the top: counts by status, total tickets, provider/model, and the token and cost breakdown for the loaded file.
- Export the loaded file anywhere via the native save dialog.
Contributions are welcome, but please note:
Open an Issue before opening a Pull Request. Discuss the bug or feature in a GitHub Issue first so we can agree on the approach. PRs without an associated Issue will be closed without review.
To work on the project locally, see Build and run from source. Before submitting, please make sure npm run typecheck and npm test pass.