AI Providers

Anthropic (Claude)

Configure Claude models in BraceKit.

3 min read

Anthropic (Claude)

Anthropic’s Claude models are known for nuanced understanding, careful reasoning, and strong safety practices.

Setup

1. Get an API Key

  1. Go to console.anthropic.com
  2. Sign in or create an account
  3. Navigate to API Keys
  4. Create a new key

2. Configure in BraceKit

  1. Open Settings
  2. Select the AI Provider tab
  3. Click Anthropic in the provider grid
  4. Paste your API key
  5. Select a model

Settings are saved automatically as you type.

Available Models

ModelAPI IDBest ForContext
Claude Opus 4.6claude-opus-4-6Most complex tasks, coding200K / 1M
Claude Sonnet 4.6claude-sonnet-4-6Best balance of speed and intelligence200K / 1M
Claude Haiku 4.5claude-haiku-4-5-20251001Fast responses, high volume200K
Claude Sonnet 4claude-sonnet-4-20250514Previous generation200K
Claude 3.5 Sonnetclaude-3-5-sonnet-20241022Legacy200K
Claude 3.5 Haikuclaude-3-5-haiku-20241022Legacy, fast200K

Note: Claude 3 Opus was retired in January 2026. Use Claude Opus 4.6 instead.

Features

Extended Thinking

Claude supports extended thinking, which shows the model’s reasoning process:

  1. Click the brain icon (🧠) in the toolbar
  2. Send your message
  3. Claude shows its thinking process
┌─────────────────────────────────────┐
│ 🧠 Thinking...                    ▾ │
├─────────────────────────────────────┤
│ The user is asking about async/    │
│ await. I should cover:             │
│ 1. What it is                      │
│ 2. How it works with Promises      │
│ 3. Common patterns and pitfalls    │
└─────────────────────────────────────┘

Async/await is syntactic sugar...

All current Claude models support extended thinking. Claude Opus 4.6 and Sonnet 4.6 also support adaptive thinking.

Vision

Claude models support image analysis:

  1. Attach an image to your message
  2. Ask a question about it
  3. Claude analyzes and responds

Function Calling

Claude supports tool calling for:

  • MCP server tools
  • Built-in tools (Google Search)

Large Context

Claude models support up to 200K tokens (1M in beta for Opus/Sonnet 4.6):

  • Analyze entire codebases
  • Process long documents
  • Extended conversations

Extended Thinking Settings

Extended thinking uses additional tokens for reasoning:

SettingEffect
EnabledShows thinking process, uses more tokens
DisabledStandard responses, fewer tokens

Note: Extended thinking uses more tokens and costs more. The thinking tokens are counted in your API usage.

Model Parameters

Configure in Settings → AI Provider under the Advanced section:

ParameterRangeEffect
Temperature0-1Higher = more creative
Max Tokens1-128KMaximum response length
Use CaseTemperatureMax Tokens
Code generation0.34096
General chat0.7Default
Analysis0.58192
Creative writing0.98192

Pricing

Anthropic charges per token. Approximate costs:

ModelInput (per 1M)Output (per 1M)
Claude Opus 4.6$5.00$25.00
Claude Sonnet 4.6$3.00$15.00
Claude Haiku 4.5$1.00$5.00

Note: Extended thinking uses additional tokens. Check Anthropic pricing for current rates.

Troubleshooting

“Overloaded”

Claude sometimes returns “overloaded” errors during peak times:

  • Wait a moment and retry
  • Try a different model (Haiku is more available)

“Context length exceeded”

  • Your conversation is too long
  • Use /compact to summarize
  • Start a new conversation

Extended thinking not showing

  • Ensure the brain icon is toggled on
  • Check you’re using a Claude model
  • Some queries may not trigger extended thinking

Model not available

  • Verify your API key has access to the model
  • Check if the model ID is correct
  • Some models may require special access