Claude Provider

Claude by Anthropic is one of the most capable AI models for code generation, reasoning, and long-context tasks.

Get an API Key

Go to console.anthropic.com
Create an account or sign in
Navigate to API Keys in the left sidebar
Click Create Key and copy it

Your key will look like: sk-ant-api03-...

Configure VibeCody

Option 1: Environment variable (recommended)

export ANTHROPIC_API_KEY="sk-ant-api03-..."
vibecli --provider claude

Option 2: Config file (~/.vibecli/config.toml)

[claude]
enabled = true
api_key = "sk-ant-api03-..."
model = "claude-sonnet-4-6"

Option 3: API key helper (for key rotation / vault integration)

[claude]
enabled = true
api_key_helper = "~/.vibecli/get-key.sh claude"
model = "claude-sonnet-4-6"

The helper script must print the key to stdout and exit 0.

Model Selection

Model	Strengths	Context	Best for
`claude-opus-4-6`	Highest quality reasoning, complex tasks	1M tokens	Architecture decisions, hard bugs, nuanced refactoring
`claude-sonnet-4-6`	Strong balance of quality and speed	200K tokens	Daily coding, code review, generation
`claude-haiku-4-5`	Fast and affordable	200K tokens	Quick questions, completions, simple edits

Default: claude-sonnet-4-6

Override from the CLI:

vibecli --provider claude --model claude-opus-4-6

Extended Thinking

Enable Claude’s extended thinking mode for complex reasoning tasks. The model will “think” internally before responding, producing higher-quality outputs for difficult problems.

[claude]
enabled = true
model = "claude-sonnet-4-6"
thinking_budget_tokens = 10000

Extended thinking uses additional tokens (and cost), but significantly improves performance on:

Multi-step debugging
Architecture design
Complex refactoring
Security analysis

Pricing

Pricing as of early 2026 (check anthropic.com/pricing for current rates):

Model	Input (per 1M tokens)	Output (per 1M tokens)
Opus 4.6	$15.00	$75.00
Sonnet 4.6	$3.00	$15.00
Haiku 4.5	$0.25	$1.25

Tip: Use Haiku for quick questions and Sonnet for coding tasks to manage costs. Reserve Opus for complex problems.

Custom API URL

For proxies or enterprise endpoints:

[claude]
enabled = true
api_url = "https://my-proxy.example.com/v1"
model = "claude-sonnet-4-6"

Verify Connection

vibecli --provider claude -c "Say hello and identify yourself"

Expected output should mention Claude by name.

Troubleshooting

Invalid API key

Error: 401 Unauthorized

Verify the key starts with sk-ant-
Check that the key has not been revoked in console.anthropic.com
If using an env var, confirm it is exported: echo $ANTHROPIC_API_KEY

Rate limited

Error: 429 Too Many Requests

Anthropic applies per-key rate limits based on your usage tier
New accounts start at a lower tier; limits increase with usage
Add a billing method to increase your tier

Context length exceeded

Error: 400 - max tokens exceeded

Use a model with a larger context window (Opus 4.6 supports 1M tokens)
Reduce the conversation history or input size
VibeCody automatically prunes old context when the window fills up

Timeout

Error: request timed out

Long prompts with extended thinking may take 30-60 seconds
Check your network connection
If behind a corporate proxy, configure api_url appropriately

Provider: Claude