Client Integrations (Claude Code / Codex CLI / SDKs)

Point existing AI clients to Inference Space by switching only the base URL and API key. The data plane is https://ai.inf.space, and authentication uses a gateway API key beginning with gk_. This page provides one-click installation scripts for Claude Code and Codex CLI, plus manual configuration for the OpenAI and Anthropic SDKs.

The private deployment integration is identical; replace ai.inf.space with your own data-plane domain → Enterprise Private Deployment.

Codex CLI

The one-click installation script writes a named provider section, [model_providers.tos-run], to ~/.codex/config.toml and stores the API key in the separate TOS_API_KEY environment variable.

curl -fsSL https://tos.run/install/codex.sh | bash -s -- --key gk_YOUR_KEY

irm https://tos.run/install/codex.ps1 -OutFile setup-codex.ps1
.\setup-codex.ps1 -Key gk_YOUR_KEY

The ~/.codex/config.toml written by the script is equivalent to:

model_provider = "tos-run"
model = "gpt-5.6-sol"
model_reasoning_effort = "medium"
preferred_auth_method = "apikey"

[model_providers.tos-run]
name = "Inference Space"
base_url = "https://ai.inf.space/v1"
env_key = "TOS_API_KEY"
wire_api = "responses"
request_max_retries = 0
stream_max_retries = 0

Key points:

• base_url = https://ai.inf.space/v1, and wire_api = responses (the Codex Responses protocol).
• env_key = TOS_API_KEY: Codex reads the API key from this environment variable. The script adds it to ~/.zshrc or ~/.bashrc.
• model_provider = tos-run: switches the default provider to the named gateway section. Otherwise Codex routes to the built-in openai provider and returns a 401.
• Use the model shown as available in the console. We recommend managing it as a configuration value.

If Codex is not installed yet: npm install -g @openai/codex.

Claude Code

The one-click installation script writes an env block to ~/.claude/settings.json and points Claude Code to the gateway's Anthropic-compatible endpoint.

curl -fsSL https://tos.run/install/claude.sh | bash -s -- --key gk_YOUR_KEY

The ~/.claude/settings.json written by the script is equivalent to the following. Existing keys are preserved and merged:

{
  "env": {
    "ANTHROPIC_BASE_URL": "https://ai.inf.space",
    "ANTHROPIC_AUTH_TOKEN": "gk_YOUR_KEY"
  }
}

Key points:

• ANTHROPIC_BASE_URL = https://ai.inf.space (the data plane, without /v1; Claude Code appends /v1/messages itself).
• ANTHROPIC_AUTH_TOKEN = gk_YOUR_KEY: sends the key to the gateway as Authorization: Bearer.

If Claude Code is not installed yet: npm install -g @anthropic-ai/claude-code.

OpenAI SDK

For the OpenAI-compatible API, set base_url to https://ai.inf.space/v1 and use a gk_ gateway key.

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://ai.inf.space/v1",
    api_key=os.environ["TOS_API_KEY"],  # Gateway key beginning with gk_
)

resp = client.chat.completions.create(
    model="qwen-plus",        # See the console for available models
    messages=[{"role": "user", "content": "Introduce yourself in one sentence."}],
)
print(resp.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://ai.inf.space/v1",
  apiKey: process.env.TOS_API_KEY, // Gateway key beginning with gk_
});

const resp = await client.chat.completions.create({
  model: "qwen-plus", // See the console for available models
  messages: [{ role: "user", content: "Introduce yourself in one sentence." }],
});
console.log(resp.choices[0].message.content);

Anthropic SDK

For the Anthropic-native Messages API, set base_url to https://ai.inf.space; the SDK appends /v1/messages. The gateway accepts Authorization: Bearer gk_... on the data plane, so pass the gateway key through auth_token.

import os
from anthropic import Anthropic

client = Anthropic(
    base_url="https://ai.inf.space",
    auth_token=os.environ["TOS_API_KEY"],  # → Authorization: Bearer gk_...
)

msg = client.messages.create(
    model="claude-sonnet-4-x",  # See the console for available models
    max_tokens=1024,
    messages=[{"role": "user", "content": "Introduce yourself in three sentences."}],
)
print(msg.content)

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  baseURL: "https://ai.inf.space",
  authToken: process.env.TOS_API_KEY, // → Authorization: Bearer gk_...
});

const msg = await client.messages.create({
  model: "claude-sonnet-4-x", // See the console for available models
  max_tokens: 1024,
  messages: [{ role: "user", content: "Introduce yourself in three sentences." }],
});
console.log(msg.content);