Welcome

Welcome to Nexus! This guide will help you start using Nexus quickly.

Nexus is your organization’s secure gateway for accessing AI models. Route all AI requests through one control point to track costs, enforce security policies, and maintain audit logs. All without changing how your applications work. Your applications keep using the same OpenAI or Anthropic code. Just change the base URL to your Nexus instance and use your Nexus API key. Everything else stays the same.

Explore by Role

Product Owner

Understand what Nexus can do for your organization: cost tracking, security policies, and compliance.

Explore →

Developer

Install Nexus, connect your applications, and start making requests through the gateway.

Explore →

Manager

Add team members, assign roles, and manage access across your organization’s projects.

Explore →

Administrator

Deploy Nexus, configure single sign-on (SSO), and manage platform-wide settings.

Explore →

Why Nexus?

Most API gateways are built for general web traffic. Nexus is purpose-built for organizations running AI under regulatory oversight: enterprises in healthcare, finance, and critical infrastructure, as well as government programs that need to follow strict rules about AI and data security.

Nexus creates a permanent, tamper-proof record of every request. If something goes wrong with a security rule, the system blocks the request instead of letting it through. Classification labels on your data ensure sensitive information is handled according to policy. These built-in controls give organizations the compliance and security they need, without having to retrofit protections onto tools that were not designed for it.

Core Features

AI Gateway

Reach OpenAI and Anthropic through one OpenAI-compatible API. Nexus supports chat completions, responses, embeddings, images, files, and vector stores. Your existing OpenAI client libraries work without changes: just change the base URL and API key.

Native Anthropic support: Anthropic SDKs and tools like Claude Code work by setting ANTHROPIC_BASE_URL to your Nexus instance. No translation layers or code changes required.

Learn how to make your first request

Cost Tracking and Budgets

Every request is priced based on the provider, model, service tier, and context band using a live model registry. You can see exactly how much each team, project, or application is spending on AI. Set budget caps at the workspace, organization, or project level to prevent runaway costs. When you hit a limit, Nexus stops accepting requests until the budget resets or is increased.

Learn more about budgets and spend tracking

Routing and Fallbacks

Specify models using a simple grammar like openai/gpt-4 or anthropic/claude-3-5-sonnet-20241022. Nexus automatically handles fallbacks: if your first choice is unavailable or rate-limited, it tries the next model in your list. This improves reliability without requiring custom retry logic in your application code.

Bring Your Own Keys (BYOK): Register your own OpenAI or Anthropic API keys in Nexus. Keys are encrypted at rest and never leave your deployment. The provider bills you directly.

Pass-Through Billing (PTB): Use shared credentials managed by your organization. Costs are deducted from a wallet, and your administrator handles billing. Nexus tries BYOK first and falls back to PTB automatically.

Learn more about model routing and fallbacks

Data Classification and Security

Label every request with a sensitivity level: public, internal, confidential, or restricted. Your organization defines the classification vocabulary and policies. Nexus enforces classification ceilings: if a project is only approved to handle internal data, requests labeled confidential are rejected before reaching the AI provider.

Classification controls which providers can handle sensitive data, how requests are logged, and what redaction rules apply to audit records. This ensures that sensitive information is handled according to your organization’s compliance requirements.

Learn more about classification

Observability and Agent Tracing

Nexus records a complete history of what your AI systems did. View requests, sessions, and agent execution traces in the console, or export them to your own observability tools using webhooks or OpenTelemetry (OTLP) collectors. This gives you the evidence you need for incident reviews, security investigations, and regulator inquiries.

Sessions: Group related requests together to track multi-turn conversations or agentic workflows. See the full execution tree showing prompts, tool calls, retries, and final responses.

Trace capture: Optionally store full request and response payloads for debugging and compliance reviews. Enable tracing per-request, per-API key, or per-project.

Learn more about sessions and tracing

Policy Engine

A dedicated policy decision point evaluates every request before it reaches an AI provider. Policies can enforce classification ceilings, apply redaction obligations, set provider routing constraints, add rate limits, or block requests entirely.

The policy engine fails closed: if a required obligation cannot be applied, the request is denied. Every policy decision emits its own audit event recording which policy decided, what it decided, and why. This ensures that policy enforcement is transparent and verifiable.

Learn more about the policy decision contract

Tamper-Evident Audit

Nexus creates a permanent, tamper-proof audit trail using per-tenant Merkle chains with signed batch roots and write-once-read-many (WORM) archival. Every request, policy decision, and access event is recorded in an immutable log that you can verify offline using cryptographic inclusion proofs.

This audit system meets the evidentiary requirements for regulated and government deployments, with controls mapped to NIST Special Publication 800-53 audit standards.

Learn more about tamper-evident audit

Multi-Tenant Identity and Access Control

Nexus organizes access in three levels:

• Workspace: Your top-level account (usually one per company)
• Organization: A team or department (like Engineering, Product, Marketing)
• Project: A specific application or use case (like a chatbot or analytics tool)

Each project gets its own API keys, usage tracking, and cost reporting. API keys are scoped to projects, so a compromised key cannot access data from other projects. This isolation limits the blast radius of security incidents.

Three roles control access:

• User: Self-service member who can create API keys and view usage for their projects
• Auditor: Can view audit logs (read-only access to compliance data)
• Manager: Full administrative control over organizations, projects, members, and settings

Authentication: Sign in with email/password or configure single sign-on (SSO) using OpenID Connect (OIDC) providers like Okta, Google Workspace, or Azure AD.

Learn more about workspace structure

Built on Open Source

Nexus is built as Rust microservices with a Next.js console, running on a fully open-source data plane:

• PostgreSQL: Relational data (users, projects, API keys)
• ClickHouse: Analytics and usage tracking
• S3-compatible object storage: Trace payloads and audit archives
• NATS: Messaging and event streaming
• Redis-compatible cache: Rate limiting

You can deploy Nexus in your own infrastructure and maintain full control over your data. No vendor lock-in, no proprietary databases, no closed-source dependencies.

Learn more about our license policy

Who is Nexus For?

Nexus is designed for organizations that need more than basic API access to AI models. It is ideal for:

• Regulated industries: Healthcare (HIPAA), finance (SOC 2, PCI-DSS), government (FedRAMP), legal
• Enterprises with compliance requirements: Organizations that need audit trails, data classification, and policy enforcement
• Teams managing multiple AI applications: Track costs, enforce budgets, and monitor usage across projects
• Security-conscious organizations: Tamper-evident logs, classification enforcement, and access controls

If you need to answer questions like “How much did we spend on AI last quarter?” or “Can we prove that confidential customer data was handled correctly?” or “Which team is using the most expensive models?”, Nexus gives you the visibility and control to find those answers.

Next Steps

• Get started: Learn how to access Nexus, create API keys, and make your first request
• Using Nexus: Understand classification, sessions, model selection, and budgets
• Installation guide: Deploy Nexus in your own infrastructure
• Developer documentation: Technical integration guides and SDK examples