Providers
Provider configuration is managed in Settings → Providers. Open Settings by clicking the gear icon in the sidebar, then select Providers.
The Providers View
Section titled “The Providers View”The Providers page has a two-panel layout:
- Left panel — list of all providers. Providers with a configured API key are shown first with a checkmark (✓). Providers without a key are shown below.
- Right panel — detail view for the selected provider: API key input, credit balance (where applicable), and model list.
Click any provider in the left panel to manage it.
Adding an API Key
Section titled “Adding an API Key”- Open Settings → Providers.
- Click a provider in the left panel.
- In the right panel, paste your API key into the input field.
- Click the eye icon to toggle show/hide.
- Navigate away — keys are saved automatically.
Keys are saved locally in the Tarsk database on your machine. They are never sent anywhere except directly to the provider’s API.
Where to Get Keys
Section titled “Where to Get Keys”| Provider | Key URL |
|---|---|
| Anthropic | console.anthropic.com/keys |
| OpenAI | platform.openai.com/api-keys |
| aistudio.google.com | |
| xAI | console.x.ai |
| DeepSeek | platform.deepseek.com |
| OpenRouter | openrouter.ai/keys |
| AIHubMix | aihubmix.com |
| Groq | console.groq.com/keys |
See Providers Overview for all providers and their key URLs.
Checking Your Credit Balance
Section titled “Checking Your Credit Balance”For OpenRouter and AIHubMix, the right panel shows your current account balance beneath the API key field. This is fetched live from the provider. If the balance reads $0.00 and you expect more, verify the key is correct.
Enabling and Disabling Models
Section titled “Enabling and Disabling Models”After adding a key, the right panel shows all available models for that provider. Each model has a checkbox:
- Checked — model appears in the model selector dropdown in every thread
- Unchecked — model is hidden from the selector
By default, no models are enabled until you check them. Enable the models you plan to use to keep the selector uncluttered.
Sorting and Filtering Models
Section titled “Sorting and Filtering Models”Use the controls above the model list:
| Filter | Shows |
|---|---|
| All | Every available model |
| Free | Models with no token cost (on supporting providers like OpenRouter) |
| Tools | Models that support tool/function calling (required for the AI agent to use its built-in tools) |
| Sort | Order |
|---|---|
| A–Z | Alphabetical by model name |
| Price ↑ | Cheapest per-token first |
| Price ↓ | Most expensive per-token first |
Onboarding Wizard
Section titled “Onboarding Wizard”When you first launch Tarsk, the Onboarding Wizard guides you through the minimum setup:
- Git Check — verifies
gitis on your PATH - Provider Setup — select a provider and enter an API key
- Model Selection — enable the models you want to use
The wizard can be reset at any time by triple-clicking the General label in the Settings sidebar. This is useful if you want to walk through it again after adding a new provider.
Using Multiple Providers
Section titled “Using Multiple Providers”You can configure as many providers as you like. Each thread has an independent model selector, so you can use Anthropic in one thread and Google in another simultaneously.
A common pattern:
- Set a capable, tool-using model (e.g. Claude Sonnet 4, GPT-4.1) as the default for active coding threads
- Set a fast, cheap model (e.g. Gemini 2.5 Flash, DeepSeek V3) for lighter tasks like summarisation or commit message generation
Local Provider
Section titled “Local Provider”Tarsk includes a Local provider for connecting to local inference servers (e.g. LM Studio, Ollama-compatible endpoints). Configure the API URL in the provider settings or via the LOCAL_API_URL environment variable. The URL is resolved dynamically when you connect.
Model Aliases
Section titled “Model Aliases”Each project can define model aliases that map role names to specific models. This lets you and the agent refer to models by purpose instead of by name.
| Alias | Use case |
|---|---|
fast | Quick responses for lightweight tasks |
local | On-device or local inference |
cheap | Lower-cost models for routine work |
smart | Most capable models for complex tasks |
Configure aliases in Project Settings under the model aliases section. Each alias maps to a specific provider + model pair. You must enable the target model in Providers first.
Once configured, you can type fast or cheap in the model selector dropdown instead of searching for the full model name. Subagents also use these aliases when their AGENT.md specifies a model alias.
Usage Cost Display
Section titled “Usage Cost Display”Completed chat messages show token usage and estimated cost in the message metadata. This helps you track spending per conversation. If your Tarsk prompt balance is empty, see Account before sending messages.
Empty Balance
Section titled “Empty Balance”An empty prompt balance blocks new chat messages until you purchase more prompts or upgrade your plan.