Quickstart: Chat in your IDE
Install the extension, activate a license, add a provider key, and send your first message — usually under five minutes.
-
Install the extension
Visual Studio 2022 (17.x) and Visual Studio 2026: install Tempr from the Visual Studio Marketplace (Extensions → Manage Extensions → search "Tempr"). VS Code: search "Tempr" in the Extensions view, or install from the Visual Studio Code Marketplace. JetBrains IDEs (IntelliJ IDEA, PyCharm, WebStorm, Rider, and other IntelliJ-Platform IDEs): install Tempr from the JetBrains Marketplace in Settings → Plugins → Marketplace.
-
Activate a license
Subscribe to Standard or Pro at temprhq.io — a license key is emailed automatically after checkout (including during the 7-day free trial). Open the Tempr tool window and paste the key in when prompted, or from Settings → License.
-
Add a provider API key
Open Settings → Providers and add a key for any provider on the supported list — a first-party provider (OpenAI, Anthropic, Google, xAI) or an aggregator (OpenRouter and others). Keys are encrypted at rest and never logged in plaintext.
-
Chat with any model
Open the Tempr tool window, pick a model from the dropdown, and start chatting. Type
@to see the list of built-in agent personas, or just describe what you want and let the default agent handle it.
Choosing an agent mode
Before you start a multi-step task, pick how much autonomy the agent has:
| Mode | Behavior |
|---|---|
| Interactive | Approve every file edit and command before it runs. Best for unfamiliar code or high-stakes changes. |
| Bypass | Edits apply automatically; whitelisted commands still run without a prompt. Good default for routine work. |
| Autopilot | The agent runs a full multi-step task end-to-end with minimal interruption. Best once you trust the task scope. |
See Agent modes & personas for the full breakdown, including named personas and multi-agent delegation.
Slash commands
Inside any chat message:
| Command | Does |
|---|---|
/explain | Explain the selected code or file. |
/fix | Diagnose and fix the selected issue. |
/tests | Generate tests for the selected code. |
/doc | Write documentation for the selected code. |
Sessions, tool preferences, and custom personas sync to the cloud on Standard and Pro, so your setup follows you across machines. The agent loop itself also runs server-side, so it survives a dropped connection instead of losing progress mid-turn.
Next steps
- Providers & BYOK — the full supported-provider list and how key encryption works.
- Agent modes & personas — named personas like
@debuggerand multi-agent delegation. - MCP servers — extend the agent with your own tools and data sources.