Key facts
| What it is | Native desktop chat client for Plugsky's API |
| Platforms | macOS, Linux and Windows builds documented on the downloads page |
| Authentication | Your own Plugsky API key, scoped and revocable from the dashboard |
| Models | 30+ models selectable in-app from one account |
| API | OpenAI-compatible endpoint shared with the SDKs and CLI |
| Free tier | Free plan with 2 free AI models, no card required |
| Governance | Same key controls, RBAC and audit logs as other API clients |
| Deployment | Can target private Plugsky deployments for restricted networks |
TL;DR
- One native client for macOS, Linux and Windows.
- Bring your own API key — same account as the API.
- Switch models per conversation from a 30+ model catalogue.
- History and context stay under your account controls.
- Useful for evaluation, support triage and internal Q&A.
How it works, step by step
- Create a Plugsky account and generate an API key in the dashboard.
- Download the build for your platform from the downloads page and install it.
- Paste the API key into the app and confirm the connection.
- Set default models for quick chat and for heavier reasoning tasks.
- Review privacy settings for conversation storage before using sensitive data.
- Revoke or rotate the key from the dashboard when a device is retired.
Try it yourself
Open the AI prompt generator →
What Plugsky Desktop is
A desktop chat application that talks to Plugsky's OpenAI-compatible API using an API key you control. It is not a separate subscription: the same account, keys, usage accounting and access controls apply as with the HTTP API. The point is convenience — a native window, keyboard-first navigation, and no browser tab or terminal needed for everyday model interaction.
Why use a desktop client instead of a browser
- Model choice: switch between 30+ models without managing provider accounts.
- Key custody: the app authenticates with a scoped key you can revoke at any time.
- Machine-level storage: conversation history stays on the device rather than a third-party web session.
- Platform coverage: one product on macOS, Linux and Windows for mixed teams.
- Consistency: behavior matches the API, so what you test in the app matches what you ship.
Team and enterprise use
Teams typically use Desktop for evaluation and internal Q&A before building integrations: compare models on real prompts, then move the winning configuration into code. Because keys are scoped and audited, IT can distribute access without sharing a shared credential. For restricted networks, configure the client to reach a private Plugsky deployment — VPC, on-prem or air-gapped — so prompts never traverse the public internet.
What to check before rollout
- Key scoping: one key per person or device, with least privilege.
- Data handling: which conversations are stored locally and for how long.
- Model policy: which aliases are approved for which data classifications.
- Updates: how the client is patched across platforms.
- Fallback: the documented API path for tasks the app does not cover.
None of these are blockers; they are the checks that make a desktop client safe to standardize on.
Honest comparison
| Option | Plugsky Desktop | Browser chat UI | Custom internal app |
|---|---|---|---|
| Platforms | macOS, Linux, Windows | Any browser | Whatever you build |
| Setup effort | Install and paste a key | Sign in | Weeks of development |
| Model switching | Built-in across 30+ models | Built-in | You implement it |
| Data location | On your machine | Provider session | Your infrastructure |
| API parity | Same endpoint as SDKs | Varies | Your integration |
Frequently asked questions
Is Plugsky Desktop free?
The client uses your Plugsky account. The free plan includes 2 free AI models with no card, and paid self-serve plans are flat monthly. Check the live pricing page for current details.
Which operating systems are supported?
macOS, Linux and Windows. Installers and the current version for each platform are listed on the downloads page.
Does Desktop use a different API from the SDKs?
No. It connects to the same OpenAI-compatible endpoint, so model behavior and account controls are consistent across clients.
Where are conversations stored?
Review the app's privacy settings for local storage behavior. Because the client runs on your device and authenticates directly, history handling is under your account and device controls.
Can it connect to a private deployment?
Yes. Point the client at a private Plugsky endpoint — VPC, on-prem or air-gapped — when policy prohibits public internet access.
How do I revoke access for a lost device?
Delete the API key used by that device in the dashboard. Scoped per-device keys make revocation immediate and contained.