Visit our [download page](https://prompten.io/download) and grab the installer for your OS: macOS (.dmg, Apple Silicon + Intel) or Windows (.msi, x64). The download is free; you only need a Promten account to sign in.
macOS: Drag Promten to Applications. Windows: Run the .msi. On first launch, sign in via Google, Microsoft, or email. The desktop app syncs your session seamlessly.
From Settings → Providers, add your desired LLM and media-gen vendors like OpenAI, Anthropic, Google Gemini, Fal, and Replicate. Paste your API keys—they're encrypted at rest.
In Models, select the LLMs and media generators you want to use from your connected providers. These become your project library, allowing you to set optional defaults for quicker access.
Open Prompt Studio to type your prompt, define placeholders (e.g., {{articleText}}), set output formats (plain text or JSON), and select models from your library. Configure per-run parameters like temperature and max tokens.
Hit 'Run all' to dispatch your prompt to all selected models concurrently. Watch their progress live in parallel columns, with errors isolated to individual models.
See outputs, costs, tokens used, and latency for each model in a single view. Instantly compare performance, quality, and price to make informed decisions.
Click 'Save scenario' to capture your prompt, models, and parameters. Re-run it later, share it with teammates, or promote it to the Prompt Catalog for production use.
