# Nexus

> Documentation for Nexus, version 1.x.
> Every entry links to the page's markdown source.

Repository: https://github.com/xraph/nexus
Site: https://www.xraph.com/docs/nexus

## Pages

- [Go Packages](https://www.xraph.com/docs/nexus/api-reference/go-packages.md): Package reference for all Nexus Go packages.
- [HTTP API](https://www.xraph.com/docs/nexus/api-reference/http-api.md): OpenAI-compatible endpoints and admin API reference.
- [Streaming API Reference](https://www.xraph.com/docs/nexus/api-reference/streaming-api.md): Go types and HTTP wire formats for streamed completions.
- [Architecture](https://www.xraph.com/docs/nexus/architecture.md): How Nexus packages fit together — Gateway, Engine, Pipeline, Providers, and Plugins.
- [Configuration](https://www.xraph.com/docs/nexus/concepts/configuration.md): Configure Nexus with functional options, structs, or YAML files.
- [Entities](https://www.xraph.com/docs/nexus/concepts/entities.md): Core entity types — Tenants, API Keys, Usage Records, and TypeID identifiers.
- [Errors](https://www.xraph.com/docs/nexus/concepts/errors.md): Error values and HTTP status code mapping.
- [Identity & Auth](https://www.xraph.com/docs/nexus/concepts/identity.md): Authentication providers, API key validation, and the Authsome adapter.
- [Multi-Tenancy](https://www.xraph.com/docs/nexus/concepts/multi-tenancy.md): Tenant-scoped requests, per-tenant configuration, quotas, and usage isolation.
- [Identity (TypeID)](https://www.xraph.com/docs/nexus/concepts/type-ids.md): How Nexus uses prefix-qualified, globally unique identifiers for every entity.
- [Model Aliases](https://www.xraph.com/docs/nexus/core/aliases.md): Map virtual model names to provider targets with per-tenant overrides and weighted routing.
- [Caching](https://www.xraph.com/docs/nexus/core/caching.md): Cache LLM responses with deterministic keys and optional semantic matching.
- [Guardrails](https://www.xraph.com/docs/nexus/core/guardrails.md): Enforce content safety with PII detection, prompt injection blocking, and content filtering.
- [Providers](https://www.xraph.com/docs/nexus/core/providers.md): Register and manage LLM providers — 30 built-in providers covering all major LLM APIs.
- [Routing](https://www.xraph.com/docs/nexus/core/routing.md): Route requests across providers using priority, cost-optimized, or round-robin strategies.
- [Transforms](https://www.xraph.com/docs/nexus/core/transforms.md): Modify requests and responses with system prompts, RAG injection, anonymization, and normalization.
- [Quick Start](https://www.xraph.com/docs/nexus/getting-started.md): Install Nexus and route your first LLM request in minutes.
- [Custom Plugin](https://www.xraph.com/docs/nexus/guides/custom-plugin.md): Build a custom lifecycle plugin for Nexus.
- [Custom Store](https://www.xraph.com/docs/nexus/guides/custom-store.md): Implement the store.Store interface for custom persistence backends.
- [Forge Extension](https://www.xraph.com/docs/nexus/guides/forge-extension.md): Run Nexus as a forge.Extension with auto-discovery and DI.
- [Full Example](https://www.xraph.com/docs/nexus/guides/full-example.md): A complete multi-tenant AI gateway with all features enabled.
- [Streaming Responses](https://www.xraph.com/docs/nexus/guides/streaming.md): SSE, NDJSON, WebSocket, and the Go channel API for streamed completions.
- [Introduction](https://www.xraph.com/docs/nexus.md): Composable AI gateway library for Go.
- [Audit](https://www.xraph.com/docs/nexus/infrastructure/audit.md): Record lifecycle events for compliance and debugging with the audit_hook extension.
- [Observability](https://www.xraph.com/docs/nexus/infrastructure/observability.md): Metrics, tracing, and monitoring for your AI gateway.
- [Plugins](https://www.xraph.com/docs/nexus/infrastructure/plugins.md): Extend Nexus with 15 lifecycle hooks for audit, metrics, webhooks, and custom processing.
- [AI21](https://www.xraph.com/docs/nexus/providers/ai21.md): Use AI21's Jamba models with Nexus — OpenAI-compatible API.
- [Anthropic](https://www.xraph.com/docs/nexus/providers/anthropic.md): Use Anthropic Claude models with Nexus — vision, tools, and extended thinking.
- [Anyscale](https://www.xraph.com/docs/nexus/providers/anyscale.md): Use Anyscale's Ray-based distributed inference with Nexus.
- [Azure OpenAI](https://www.xraph.com/docs/nexus/providers/azureopenai.md): Use Azure-hosted OpenAI models with Nexus — deployment-based URL and api-key auth.
- [Amazon Bedrock](https://www.xraph.com/docs/nexus/providers/bedrock.md): Use Amazon Bedrock models with Nexus — AWS SigV4 auth and Converse API.
- [Cerebras](https://www.xraph.com/docs/nexus/providers/cerebras.md): Use Cerebras Wafer Scale Engine for ultra-fast Llama inference with Nexus.
- [Cohere](https://www.xraph.com/docs/nexus/providers/cohere.md): Use Cohere Command and Embed models with Nexus — native embeddings with input types.
- [Deepinfra](https://www.xraph.com/docs/nexus/providers/deepinfra.md): Use Deepinfra for serverless open-source model inference with Nexus.
- [DeepSeek](https://www.xraph.com/docs/nexus/providers/deepseek.md): Use DeepSeek reasoning models (R1) and chat models with Nexus.
- [Fireworks AI](https://www.xraph.com/docs/nexus/providers/fireworks.md): Use Fireworks AI for fast open-source model inference with Nexus.
- [Google Gemini](https://www.xraph.com/docs/nexus/providers/gemini.md): Use Google Gemini models with Nexus — vision, embeddings, and tool calling.
- [Groq](https://www.xraph.com/docs/nexus/providers/groq.md): Use Groq's ultra-fast LPU inference with Nexus.
- [Hyperbolic](https://www.xraph.com/docs/nexus/providers/hyperbolic.md): Use Hyperbolic for open-source model inference with Nexus.
- [Jina AI](https://www.xraph.com/docs/nexus/providers/jinaai.md): Use Jina AI embeddings with Nexus — embeddings only via OpenAI-compatible API.
- [Lepton AI](https://www.xraph.com/docs/nexus/providers/lepton.md): Use Lepton AI's serverless GPU inference with Nexus.
- [LM Studio](https://www.xraph.com/docs/nexus/providers/lmstudio.md): Run models locally with LM Studio and connect to Nexus — no API key required.
- [Mistral](https://www.xraph.com/docs/nexus/providers/mistral.md): Use Mistral models with Nexus — embeddings, vision, and code generation.
- [Nebius](https://www.xraph.com/docs/nexus/providers/nebius.md): Use Nebius for EU-region AI inference and embeddings with Nexus.
- [Novita AI](https://www.xraph.com/docs/nexus/providers/novita.md): Use Novita AI for serverless open-source model inference with Nexus.
- [NVIDIA NIM](https://www.xraph.com/docs/nexus/providers/nvidia.md): Use NVIDIA NIM for GPU-optimized inference and embeddings with Nexus.
- [Ollama](https://www.xraph.com/docs/nexus/providers/ollama.md): Run models locally with Ollama and connect to Nexus — no API key required.
- [OpenAI](https://www.xraph.com/docs/nexus/providers/openai.md): Use OpenAI models (GPT-4o, o1, embeddings) with Nexus.
- [OpenAI-Compatible (Generic)](https://www.xraph.com/docs/nexus/providers/opencompat.md): Connect any OpenAI-compatible API to Nexus with the generic opencompat provider.
- [OpenRouter](https://www.xraph.com/docs/nexus/providers/openrouter.md): Access hundreds of models through OpenRouter's unified API with Nexus.
- [Perplexity](https://www.xraph.com/docs/nexus/providers/perplexity.md): Use Perplexity's search-augmented generation models with Nexus.
- [SambaNova](https://www.xraph.com/docs/nexus/providers/sambanova.md): Use SambaNova's custom RDU silicon for AI inference with Nexus.
- [Together AI](https://www.xraph.com/docs/nexus/providers/together.md): Use Together AI for fast open-source model inference and embeddings with Nexus.
- [Google Vertex AI](https://www.xraph.com/docs/nexus/providers/vertex.md): Use Vertex AI-hosted Gemini models with Nexus — GCP OAuth2 auth with project/location routing.
- [Voyage AI](https://www.xraph.com/docs/nexus/providers/voyageai.md): Use Voyage AI's high-quality text embeddings with Nexus — embeddings only.
- [xAI (Grok)](https://www.xraph.com/docs/nexus/providers/xai.md): Use xAI Grok models with Nexus — vision and tool calling support.
- [MongoDB Store](https://www.xraph.com/docs/nexus/stores/mongo.md): MongoDB store for document-oriented and horizontally-scalable deployments.
- [PostgreSQL](https://www.xraph.com/docs/nexus/stores/postgres.md): Configure PostgreSQL as the persistence backend for tenants, API keys, and usage data.
- [SQLite Store](https://www.xraph.com/docs/nexus/stores/sqlite.md): Lightweight SQLite store for development, testing, and single-node deployments.