About ZBStream
ZBStream is a developer-focused LLM gateway: one API key and an OpenAI-compatible interface give you unified access to models from many providers, with routing, failover, billing, and usage tracking built in.
Documentation Index
Use the documentation menu on the left to browse the current Quick Start, API Guide, interface references, and changelog. Before integrating, we recommend reading First API Call, Models & Pricing, and Error Codes.
Core Principles
ZBStream connects different models and providers behind one interface, so your application can adjust model selection, channel configuration, and cost strategy without rewriting business code.
Why ZBStream?
- Price and performance: centrally manage model pricing, markups, and availability across connected channels to pick the right model for each workload.
- Standardized API: compatible with the OpenAI API and common SDKs — switch models or providers without rewriting your calling code.
- Real usage insights: the console records model requests, input tokens, cache-hit tokens, output tokens, and credit consumption.
- Unified settlement: every model and provider shares one credit account and usage ledger, with clear and auditable billing rules.
- Higher availability: configure multiple upstream channels per model and use routing policies to reduce the impact of a single provider outage.
- Agent ecosystem: works with agent tools that support a custom base URL, plus the BewareClaw desktop agent.
Note
Actual models, channels, prices, rate limits, and failover behavior are subject to the current console configuration.