Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/Emanuele-web04/dpcode/llms.txt

Use this file to discover all available pages before exploring further.

The Claude Code provider connects DP Code to Anthropic’s Claude Code CLI agent. The provider kind is claudeAgent. When you start a session, DP Code launches the claude binary as a subprocess, manages the session lifecycle, and streams events to your browser.

​
Prerequisites

1

Install Claude Code CLI

Install the Claude Code CLI from Anthropic. The recommended method is npm:
npm i -g @anthropic-ai/claude-code
After installation, verify it is available:
claude --version
2

Authenticate

Run claude in your terminal to complete the login flow and authorize against your Anthropic account or API key.
claude
3

Confirm the binary is on your PATH

DP Code resolves claude from PATH by default. If you installed it to a custom location, set binaryPath in your provider settings to the absolute path.

​
Settings

Configure the Claude Code provider under providers.claudeAgent in your DP Code server settings.
​
providers.claudeAgent.binaryPath
string
default:"claude"
Path to the claude executable. Accepts a bare binary name resolved from PATH or an absolute path.
​
providers.claudeAgent.launchArgs
string
default:""
Additional command-line arguments passed to the claude binary at startup, provided as a single string. Use this to pass custom flags that are not otherwise exposed by DP Code.Example: "--dangerously-skip-permissions"
​
providers.claudeAgent.customModels
string[]
default:"[]"
List of additional model slugs to make available in the model picker. Each entry must be a non-empty string with a maximum of 256 characters.
​
providers.claudeAgent.enabled
boolean
default:"true"
Set to false to disable the Claude Code provider. Existing sessions continue until you reload.
The launchArgs field is a raw string appended to the launch command. Separate multiple flags with spaces, as you would in a terminal.

​
Model selection

The following models are available by default. DP Code selects claude-sonnet-4-6 when no model is specified.
Model slugDisplay name
claude-opus-4-7Claude Opus 4.7
claude-opus-4-6Claude Opus 4.6
claude-opus-4-5Claude Opus 4.5
claude-sonnet-4-6Claude Sonnet 4.6
claude-haiku-4-5Claude Haiku 4.5
You can also add custom model slugs using customModels.

​
Model aliases

AliasResolves to
opusclaude-opus-4-7
opus-4.7claude-opus-4-7
opus-4.6claude-opus-4-6
opus-4.5claude-opus-4-5
sonnetclaude-sonnet-4-6
sonnet-4.6claude-sonnet-4-6
haikuclaude-haiku-4-5
haiku-4.5claude-haiku-4-5

​
Reasoning effort (extended thinking)

Opus and Sonnet models support effort levels. Select the level in the model options panel.
ValueLabel
lowLow
mediumMedium
highHigh (default for most models)
xhighExtra High
maxMax
ultrathinkUltrathink
ultrathink is a prompt-injected effort level — DP Code injects it as a reasoning instruction rather than passing it as a parameter flag. It is available on Opus 4.7, Opus 4.6, and Sonnet 4.6.

​
Fast mode

Opus 4.7 and Opus 4.6 support fast mode for reduced latency.

​
Thinking toggle

Claude Haiku 4.5 exposes an explicit thinking toggle instead of effort levels. Enable it in the model options panel to activate extended thinking.

​
Context window

Opus and Sonnet models offer two context window sizes:
ValueLabelDefault
200k200kYes
1m1M—

Claude Code documentation

Official setup guide and feature reference for Claude Code.

Anthropic model overview

Current Claude model capabilities and availability.

Build docs developers (and LLMs) love

Get started for freeTalk to us