Enterprise LLM Gateway: Governed Team LLM API Access

The ReclaimLLM Enterprise LLM Gateway is a hosted, governed proxy for team LLM API traffic. Instead of developers holding provider API keys directly, applications point at the gateway and authenticate with a team-scoped gateway key — the gateway resolves that key to your org’s provider credentials and model policy, forwards the request, and logs it back to your dashboard with full org/team attribution.

The Enterprise LLM Gateway is an enterprise-only feature. Contact the ReclaimLLM team to enable it for your organization.

Why use it over the local proxy

Do not mix these up. rclm-proxy (http://localhost:9800) is the Local API Proxy — personal capture with your own provider keys. The Enterprise Gateway is hosted at https://proxy.reclaimllm.com/<org-slug>/v1 and authenticates with a team gateway key (rclm-gw-…). Setting OPENAI_BASE_URL=http://localhost:9800 does not use the enterprise gateway.

	Local API Proxy	Enterprise LLM Gateway
Host	`http://localhost:9800` (`rclm-proxy`)	`https://proxy.reclaimllm.com/<org-slug>/v1`
Auth	Personal `RECLAIMLLM_API_KEY` + your provider keys	`Authorization: Bearer rclm-gw-…`
Provider credentials	On the developer machine	Org-managed, never exposed to developers
Policy	None (passthrough capture)	Team-scoped providers / models
Plan	Free / Paid / Enterprise	Enterprise only

The Local API Proxy runs on a developer’s own machine using their own provider keys — it’s individual capture. The Enterprise Gateway is the org-level equivalent: credentials live centrally, access is scoped per team, and every call is attributed and policy-checked before it reaches the provider.

Org-managed credentials

Admins store OpenAI, Azure OpenAI, Anthropic, and Gemini credentials once at the org level. Developers never see them.

Team-scoped keys

Each team gets its own gateway key with its own allowed providers and model policy. Revoking a key doesn’t affect other teams.

OpenAI-compatible

Point any OpenAI-compatible SDK at the gateway URL and swap in a gateway key — no other code changes.

Full attribution

Every call is logged as a session with org, team, model, tokens, and duration — visible in the same dashboard as hooks, proxy, and extension sessions.

How to use it

Once an org admin has issued you a team gateway key, point your app at the gateway host and authenticate with the key as a bearer token — no other client changes needed for an OpenAI-compatible SDK.

Base URL:  https://proxy.reclaimllm.com/<org-slug>/v1
Auth:      Authorization: Bearer <gateway-key>

Supported endpoints: /v1/chat/completions (streaming and non-streaming), /v1/responses, /v1/embeddings.

OpenAI

curl https://proxy.reclaimllm.com/acme/v1/chat/completions \
  -H "Authorization: Bearer rclm-gw-your-team-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-5.6-luna",
    "messages": [
      { "role": "user", "content": "Write a short release note." }
    ]
  }'

Moonshot

curl https://proxy.reclaimllm.com/acme/v1/chat/completions \
  -H "Authorization: Bearer rclm-gw-your-team-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "moonshot/kimi-k2.6",
    "messages": [
      { "role": "user", "content": "Compare FCF of amazon and microsoft?" }
    ]
  }'

Azure OpenAI

curl https://proxy.reclaimllm.com/acme/v1/chat/completions \
  -H "Authorization: Bearer rclm-gw-your-team-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "azure/gpt-4o-mini",
    "messages": [
      { "role": "user", "content": "Summarize this incident report." }
    ]
  }'

Embeddings

curl https://proxy.reclaimllm.com/acme/v1/embeddings \
  -H "Authorization: Bearer rclm-gw-your-team-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "azure/text-embedding-3-small",
    "input": ["Searchable context from an LLM session."]
  }'

Credential-related fields in the request body (api_key, api_base, headers, and similar) are always ignored — the gateway only ever forwards using the org’s configured provider credentials, so a gateway key can’t be used to redirect traffic to a different provider or endpoint.

Request models must use a LiteLLM provider prefix (openai/gpt-5.6-luna, not bare gpt-5.6-luna). Org allowlists store the bare name. Full contract, more providers, and SDK examples: Enterprise LLM Gateway guide.

Set up the gateway

Configuring provider credentials, creating gateway keys, per-team policy, and full request examples (chat completions, responses, embeddings, Python/TypeScript SDKs) are covered in the Enterprise LLM Gateway guide.

Get Started

Capture

User Portal

Privacy & Security

Enterprise

Plans & Billing

Enterprise LLM Gateway: Governed Team LLM API Access

Why use it over the local proxy

Org-managed credentials

Team-scoped keys

OpenAI-compatible

Full attribution

How to use it

OpenAI

Moonshot

Azure OpenAI

Embeddings

Set up the gateway

​Why use it over the local proxy

Org-managed credentials

Team-scoped keys

OpenAI-compatible

Full attribution

​How to use it

​OpenAI

​Moonshot

​Azure OpenAI

​Embeddings

​Set up the gateway

Why use it over the local proxy

How to use it

OpenAI

Moonshot

Azure OpenAI

Embeddings

Set up the gateway