My Account
BYTEBOOST API SETUP

Connect your API key.
Pick your tool.

Step-by-step setup for the clients we checked.

← Back to API key checker

Before you start

  • Use an API key issued to you by the provider. Keep it private and do not put it in a public project or source control.
  • These tested examples use the claude-fable-5.1 model selector. The API checker lists the current selectors.
  • Requests from coding tools go directly to tokify.sale. The ByteBoost key checker does not proxy those requests.
TESTED

Claude Code

A one-turn text request was completed with Claude Code using the Anthropic Messages API. Open a new terminal and set these values for that terminal session.

Windows PowerShell

$env:ANTHROPIC_BASE_URL = "https://tokify.sale"
$env:ANTHROPIC_AUTH_TOKEN = "YOUR_API_KEY"
$env:ANTHROPIC_MODEL = "claude-fable-5.1"
$env:ANTHROPIC_SMALL_FAST_MODEL = "claude-fable-5.1"
claude

macOS / Linux

export ANTHROPIC_BASE_URL="https://tokify.sale"
export ANTHROPIC_AUTH_TOKEN="YOUR_API_KEY"
export ANTHROPIC_MODEL="claude-fable-5.1"
export ANTHROPIC_SMALL_FAST_MODEL="claude-fable-5.1"
claude

Replace YOUR_API_KEY locally. Do not share a terminal transcript that includes the key.

TESTED

OpenCode

A one-turn text request was completed in OpenCode with the Anthropic provider adapter.

  1. Run /connect, choose Other, enter byteboost as the provider ID, and enter your API key.
  2. Merge this provider into the opencode.json in your project root. Keep any existing settings.
  3. Run /models and select byteboost/claude-fable-5.1.
{
  "$schema": "https://opencode.ai/config.json",
  "model": "byteboost/claude-fable-5.1",
  "provider": {
    "byteboost": {
      "npm": "@ai-sdk/anthropic",
      "name": "ByteBoost Gateway",
      "options": {
        "baseURL": "https://tokify.sale/v1"
      },
      "models": {
        "claude-fable-5.1": {
          "name": "Claude Fable 5.1"
        }
      }
    }
  }
}

The configuration contains no API key; OpenCode keeps the credential in its local authentication store.

CUSTOM MESSAGES ENDPOINT

Visual Studio Code

VS Code has a Custom Endpoint provider that accepts Anthropic Messages endpoints. This configuration path has not been end-to-end tested in the VS Code UI.

  1. Open Chat: Manage Language Models from the Command Palette.
  2. Choose Add Models → Custom Endpoint, name the provider, enter your key, and select Messages.
  3. Use the model entry below in the generated chatLanguageModels.json. Keep the key in VS Code's secure input, not in this file.
[
  {
    "name": "ByteBoost Gateway",
    "vendor": "customendpoint",
    "apiKey": "${input:byteboostApiKey}",
    "apiType": "messages",
    "models": [
      {
        "id": "claude-fable-5.1",
        "name": "Claude Fable 5.1",
        "url": "https://tokify.sale/v1/messages"
      }
    ]
  }
]
NOT CONFIRMED

Hermes Agent

Our Hermes test request returned HTTP 402 even though the same key checker showed quota remaining. That mismatch needs investigation with the gateway provider before Hermes can be recommended. Avoid repeated retries, which may still count as requests.

NOT TESTED

Claude Desktop

Claude Desktop was not installed on the test PC, and its standard chat sign-in is not a custom model endpoint setting. Use Claude Code or a client with a Custom Endpoint option to connect this API.

We verified short text responses only. Tool use, streaming, vision, long-context behavior, and every other model selector still need separate checks.

Check a key or view the current model list ↗