Flaq AI Agent Guide
Practical Guide for Using Claude Code CLI with Flaq AI's API Services and Open-Source AI Agent Skills
Use Flaq AI's Claude Code-compatible endpoint to quickly configure Claude Code CLI with Flaq and access our model services from the terminal. After configuring your API key, follow the examples in this guide to choose a currently supported model and verify the setup. You can also explore reusable open-source Claude Code skills to extend AI agent capabilities and support additional workflows.
Install
Install Claude Code CLI
Install the Claude Code CLI for your system, then confirm the claude command works before adding Flaq AI credentials.
Install Claude Code CLI
Choose the installer that matches your system, then run the install command below.
Official install script
Verify Claude Code CLI
Confirm the claude command is available before adding Flaq AI credentials.
Verify installation
Install Claude Code CLI
Choose the installer that matches your system, then run the install command below.
Windows PowerShell
Verify Claude Code CLI
Confirm the claude command is available before adding Flaq AI credentials.
Verify installation
Configure
Connect Claude Code CLI to Flaq AI
Choose a model currently supported by the Flaq Claude Code endpoint, add your Flaq API key as an Anthropic auth token, and point Claude Code CLI to the Flaq API base URL. Open-source Claude Code skills cover reusable workflows for coding, deployment, testing, documents, creative work, and other tasks.
Balanced Claude model for everyday coding workflows, refactors, and terminal assistance.
Recommended Claude model for stronger reasoning, architecture work, and complex code changes.
Flagship Claude option for demanding coding sessions and deeper analysis.
Use temporary environment variables in the current terminal session. This is the fastest way to try Flaq. It does not change ~/.claude/settings.json or your shell profile.
- These exports apply only to the current terminal session.
- They do not modify
~/.claude/settings.jsonor your shell profile.
Set Flaq environment for this terminal
For a one-off run, you can skip ANTHROPIC_MODEL and use claude --model claude-sonnet-4-6 instead.
- These exports apply only to the current terminal session.
- They do not modify
~/.claude/settings.jsonor your shell profile.
Set Flaq environment for this terminal
For a one-off run, you can skip ANTHROPIC_MODEL and use claude --model claude-opus-4-6 instead.
- These exports apply only to the current terminal session.
- They do not modify
~/.claude/settings.jsonor your shell profile.
Set Flaq environment for this terminal
For a one-off run, you can skip ANTHROPIC_MODEL and use claude --model claude-opus-4-7 instead.
Use temporary environment variables in the current terminal session. This is the fastest way to try Flaq. It does not change ~/.claude/settings.json or your shell profile.
- These exports apply only to the current terminal session.
- They do not modify
~/.claude/settings.jsonor your shell profile.
Set Flaq environment for this terminal
For a one-off run, you can skip ANTHROPIC_MODEL and use claude --model claude-sonnet-4-6 instead.
- These exports apply only to the current terminal session.
- They do not modify
~/.claude/settings.jsonor your shell profile.
Set Flaq environment for this terminal
For a one-off run, you can skip ANTHROPIC_MODEL and use claude --model claude-opus-4-6 instead.
- These exports apply only to the current terminal session.
- They do not modify
~/.claude/settings.jsonor your shell profile.
Set Flaq environment for this terminal
For a one-off run, you can skip ANTHROPIC_MODEL and use claude --model claude-opus-4-7 instead.
Verify
Verify Flaq in Claude Code CLI
Run a simple prompt from your terminal. A normal response means Claude Code CLI can call models through Flaq AI.
Start interactive mode
Start interactive mode
Start interactive mode
Quick test command
Quick test command
Quick test command
Run diagnostics
Run diagnostics
Run diagnostics
Start interactive mode
Start interactive mode
Start interactive mode
Quick test command
Quick test command
Quick test command
Run diagnostics
Run diagnostics
Run diagnostics
Skills
Extend Claude Code with Open-Source AI Agent Skills
Claude Code skills are reusable packages of instructions, workflows, scripts, and reference files that extend AI agent capabilities across recurring tasks. Explore skills for coding, deployment, testing, documents, creative work, and other workflows without repeating the same instructions in every session. Review each skill's source, required tools, and permissions before using it.
Browse this multilingual repository for categorized Claude Code skill links, usage notes, source references, and installation details.
Troubleshooting