Getting Started

Introduction

Spineforge is the audit-ready AI agent gateway: one control plane to observe, govern, secure, and manage AI agents at scale.

What is Spineforge?

Spineforge is a drop-in reverse proxy that sits between your AI agents and the LLM providers they call. Every request from an IDE assistant, a CLI, a custom agent or a cloud-hosted workload passes through the gateway, so Spineforge can attribute it to an agent, enforce policy on it, and write it to an audit trail, without you rewriting a single agent.

Without a central point of control, agents share API keys through env files and IDE settings. They can't be attributed, budgeted, scoped or revoked independently. Spineforge fixes that at the root, which unlocks everything built on top of it: cost attribution, permission control, audit trails and shadow-agent detection.

Two lines to integrate

There is no SDK to install and no architecture to change. Point your agent at the gateway and authenticate with a Spineforge gateway key:

1

Base URL: point your agent at the Spineforge gateway

https://spineforge-backend.onrender.com/gateway/v1
2

API key: use your Spineforge gateway key

sfgw-xxxx-xxxx

That's it: your agent now talks to the gateway using the standard OpenAI-compatible API. Your real provider keys (OpenAI, Anthropic, Google Gemini, Groq…) live in the Spineforge Vault and are never handed to agents.

The four pillars

  • Agent Observability & Telemetry. Every request is traced with agent, provider, model, tokens, latency and cost. Cost is computed once at write time from a rate card and stored, not recalculated on every dashboard load.
  • Enterprise Governance & Control. Per-agent scopes, budget caps and rate limits are enforced inline, and a single agent can be revoked instantly with a kill-switch.
  • Regulatory Compliance & Security. An encrypted key vault, SSO, and audit trails aligned to the standards your compliance teams care about. See the Security page for what is live and what is in progress.
  • Unified AI FinOps & Access Management. One place to see every dollar of AI spend by agent and model, and to issue, scope and rotate access for every agent and developer.

Works with the tools you already use

IDEs & Editors Cursor, VS Code (native chat and extensions), and any OpenAI-compatible editor.CLI Tools OpenCode, Antigravity CLI, and any CLI that reads a base URL and an API key.Custom Agents Python, Node, LangChain, raw HTTP: any agent that can set a base URL.Cloud-Hosted Agents Containers, Kubernetes, serverless and CI. Set two environment variables.

Who Spineforge is for

Teams deploying AI agents at scale: platform and engineering teams shipping agents and IDE assistants, security and compliance teams who need enforcement and an audit trail, and finance and AI leadership who need to see and control spend.

Current status

The gateway is live. OpenAI-compatible routing with streaming, the encrypted key vault, gateway keys with scopes, budget caps and rate limits, instant revocation, and the observability dashboard are all available today. Compliance certifications, data-handling modes, SAML, and self-hosted deployment are on the roadmap and clearly labelled as in progress.

Next steps