Integration Guides

End-to-end walkthroughs for every TokenHub model family — from getting your API key to production-ready curl, Python and Node.js snippets. All endpoints are OpenAI-compatible, so you can reuse the official SDKs with a single base_url change.

Agent

Claude Code & Agent Tools

Claude Code, OpenClaw, Cline, Roo Code

Connect Claude Code, OpenClaw, Cline, Roo Code and any Anthropic-protocol client to TokenHub via our native /v1/messages endpoint. One th- key works on both protocols; tool use and MCP all pass through losslessly.

Anthropic-NativeClaude CodeOpenClawMCP / Tool Use
Read the guide
Chat

Claude API

Anthropic Claude 3 → 4.7 series

Call Claude (Opus / Sonnet / Haiku) via our OpenAI-compatible /chat/completions endpoint. Streaming, tool use, and vision all supported. Drop-in with the official openai SDK.

OpenAI-CompatibleStreamingTool UseVision
Read the guide
Chat

DeepSeek V4

DeepSeek 2026 thinking models

Call DeepSeek V4-Pro and V4-Flash via our OpenAI-compatible /chat/completions endpoint. Thinking mode with reasoning_content, 1M context, 384K output, streaming and tool use all supported.

OpenAI-CompatibleThinking Mode1M ContextTool Use
Read the guide
Image

GPT-Image-2

OpenAI image generation & editing

Text-to-image and image-to-image editing with OpenAI's latest image model. Synchronous response, multiple sizes up to 4K, three quality levels, and multi-image composition.

Text → ImageImage → ImageUp to 4KSynchronous
Read the guide
Video

Seedance 2.0

ByteDance video generation

High-quality text-to-video and image-to-video with Seedance 2.0. First/last frame conditioning, multi-shot reference mode, and async task polling.

Text → VideoImage → VideoFirst/Last FrameAsync
Read the guide

Looking for something else?

Browse the full model catalog and pricing, or jump straight into the interactive docs.