Verify & First Use

Reload Your Shell

This page picks up right after the one-command setup. Once the setup script finishes, reload your shell profile so the new environment variables and PATH take effect:

macOS / Linux

source ~/.zshrc

If you're not using zsh, source whichever profile file the script mentioned at the end of its output.

Alternatively, open a new terminal window — this is the safest approach.

Verify the Installation

claude --version

If you see a version number, Claude Code is installed and configured.

Quick Connection Test

claude -p "Reply with exactly: TOKEN101_OK"

If everything is configured correctly, Claude will respond with TOKEN101_OK.

Verification Checklist

The setup script outputs a detailed checklist. After running the script, verify each item:

Check	Command / Action
`/model` menu shows: sonnet, opus, haiku, and a custom model	Run `/model` inside Claude Code
1M context is offered as an opt-in during setup (default is 200K)	Re-run the setup script and check the context-window prompt
`sonnet` maps to: `claude-sonnet-4-5-20250929`	Run `claude -p "what model are you?"` after selecting sonnet
`opus` maps to: `claude-opus-4-6`	Same as above, select opus
`haiku` maps to: `claude-haiku-4-5-20251001`	Same as above, select haiku
`subagent`: uses the sonnet slot by default (`claude-sonnet-4-5-20250929`)	Check subagent model in settings
Custom model maps to: `gpt-5.3-codex`	Select custom model and ask

If any check fails, see the troubleshooting section below or re-run the setup script.

Make Your First API Call

Once the connection test passes, you're ready to use the tools with Token101 as the backend.

Claude Code — Interactive Session

Navigate to any project and start using Claude Code:

cd ~/my-project
claude

Claude Code will index your project and let you ask questions, write code, and debug issues interactively.

Claude Code — One-Shot Question

claude -p "Explain what this project does"

Codex — Interactive Session

cd ~/my-project
codex

Codex — One-Shot Question

codex "Review src/auth.ts for potential bugs"

Codex — Specify a Model

codex --model gpt-5.3-codex "Write unit tests for src/auth.ts"

Verify the Backend

To confirm Token101 is handling your requests, ask either tool which model it's using:

claude -p "What model are you?"

codex "What model are you?"

The response should mention a Token101-supported model (e.g., Claude Sonnet 4.5, GPT-5.1).

Troubleshooting

401 Error — Invalid API Key

Check that the API Key is entered correctly
Check if the API Key has expired — generate a new one at token.ppthub.shop/settings/apikeys
Confirm the key starts with sk-

Models Not Showing in /model

Exit current Claude Code / CLI session
Open a new terminal window
Restart with claude or claude --dangerously-skip-permissions

ANTHROPIC_AUTH_TOKEN and ANTHROPIC_API_KEY both work on their own — Token101 accepts either. The only problem is having both set at once, because which header Claude Code sends then depends on its version. Check ~/.zshrc and ~/.zprofile and keep just one variable (the install script writes ANTHROPIC_AUTH_TOKEN). See Claude Code for the full explanation.

Further Help

Claude Code Integration — advanced configuration and detailed troubleshooting

Still Not Resolved?

If none of the steps above fixed your problem, email us at [email protected] and, where you can, include:

The rough time it happened and the model you were using
The full error response (JSON body)
The call reference from the response headers, if any

We'll get back to you as soon as we can.

What's Next?

Congratulations — you're up and running with Token101! 🎉

Here's where to go from here:

Supported Models — explore all available models
Claude Code Integration — learn advanced configuration
Streaming — how to use streaming for real-time responses
Billing & Credits — understand how usage is charged