XiuRouter

XiuRouter is a unified AI model API gateway built by XiuLab Inc that lets developers access Claude, GPT, Gemini, Grok, and other leading models through a single endpoint. It supports multiple native API protocols—OpenAI Chat Completions, OpenAI Responses, Anthropic Messages, and Gemini GenerateContent—so existing clients and SDKs can connect without rewriting request logic. The service offers two service tiers (Value and Managed) with usage-based pricing that the site claims can save 77–95%+ compared to direct provider reference prices.

What It Is

XiuRouter is an API routing layer that sits between a developer's application and multiple AI model providers. Rather than managing separate API keys, base URLs, and billing accounts for each provider, developers point their existing OpenAI-compatible SDK, Anthropic client, or Gemini client at XiuRouter's endpoint and use a single XiuRouter API key. The router then handles model selection, service tier routing, and cost tracking on the backend.

Protocol and Integration Support

XiuRouter natively supports four API formats, selectable by Base URL:

• OpenAI Responses (/v1/responses) — recommended for OpenAI SDKs, Codex, and agents
• OpenAI Chat Completions — for broad OpenAI-compatible tooling
• Anthropic Messages — for Anthropic SDK users
• Gemini GenerateContent — for Google Gemini clients

The site also lists guided setup for agent integrations, and the integrations page covers 15 applications. This multi-protocol design means migration from a single provider to XiuRouter typically requires only a Base URL and API key swap.

Service Tiers and Pricing Model

XiuRouter offers two service tiers per model:

• Value — prioritizes lower prices for everyday use; model availability and response consistency may vary
• Managed — designed for workloads requiring more consistent availability and responses

Pricing is usage-based (USD per 1M tokens). The homepage price comparison table shows example savings ranging from 77% to over 95% versus provider reference prices for models like claude-opus-5, gpt-5.6-sol, gemini-3.6-flash, and grok-4.6. The site notes that savings vary by model, tier, and price type (input vs. output vs. cache).

Developer Workflow

The setup flow is straightforward:

1. Add funds and convert wallet balance to router credit
2. Create an API key with a chosen service tier and model access
3. Set the matching Base URL in your client
4. Send a request and verify the model ID and response
5. Review usage (model, tier, tokens, cost, status) in the Usage dashboard

The console includes a Playground for testing requests with streaming output, a Usage view for per-key cost and latency inspection, and a Rankings page showing the most-used models over the last 7 days.

Observability and Monitoring

XiuRouter provides built-in observability tools: the Playground lets developers inspect streaming responses and exact model IDs before production; the Usage dashboard records request status, token counts, latency, and cost per API key; and the Rankings page tracks model share of requests over the last 7 days with movement trends. This makes it easier to reconcile spend and compare model performance without external tooling.

Platform Context

XiuRouter is one of several products under XiuLab Inc, alongside XiuStore, XiuBench, and the forthcoming XiuGEO. The router targets agent developers, teams migrating from single-provider setups, and applications that need multi-model access or cost optimization across providers. Documentation is hosted at docs.xiu.ai/en/router/ and includes a quickstart guide.