Skip to content
BrainRoad BrainRoad
Documentation Menu

Choosing a Provider

The AI providers BrainRoad supports and how to pick one when your trial credit runs out.

On this page

When your included trial credit runs out, you add your own API key from an AI provider. This page lists what BrainRoad supports and how to choose.

Supported providers

You can add keys for these providers from the API Keys tab on the Helper page:

ProviderModelsNotes
AnthropicClaudeRecommended primary choice; strong at reasoning, writing, and code. Keys start with sk-ant-. console.anthropic.com
OpenAIGPTVersatile general-purpose models. Keys start with sk-. platform.openai.com/api-keys
GoogleGeminiHas a free tier, so it’s the cheapest way to start on your own key. Keys start with AIza. aistudio.google.com/apikey
OpenRouterMany providers via one keyOne account for Claude, GPT, Gemini, and more, including some free models. This is also the service behind your trial key. Keys start with sk-or-. openrouter.ai/keys
xAIGrokThe trial’s default model family, on your own account. Keys start with xai-. console.x.ai
MistralMistralEuropean provider, good at coding and structured tasks. console.mistral.ai
MoonshotKimiStrong at vision and video tasks. platform.moonshot.cn
Kilo GatewayKilo CodeSpecialized coding models. kilocode.ai

You can configure more than one provider on the same helper.

How to choose

If you’re not sure, this is the short version:

  • Free to start: Google Gemini’s free tier, or OpenRouter’s free models.
  • Best quality: Anthropic Claude. If you only ever add one key, this is the usual pick.
  • Maximum flexibility: OpenRouter — one key, access to most major models, easy to switch models later.
  • You already have a key: just paste it. Any supported provider works.

The guided setup on the Helper page walks you through this same decision if you’d rather answer one question and get a recommendation.

What it costs

You pay the provider directly at their published rates; BrainRoad adds no markup and the subscription does not include model usage. Actual spend depends entirely on how much work your helper does, so check the provider’s own pricing page for current rates. The provider picker in the dashboard shows a rough cost estimate next to each option.

Two habits keep costs predictable:

  1. Set a spending limit in the provider’s billing dashboard before you start.
  2. Watch the Usage page (/dashboard/usage) in your BrainRoad dashboard to see token usage over time.