🛡️
AshLab Gateway
LLM Connection Request Portal
External Developer Access

Request an LLM Connection

Use this form to register your LLM provider endpoint with the AshLab Gateway. Once approved by an admin, you will receive an MCP connection string that routes all your AI traffic through the enterprise proxy — giving you governance, telemetry, cost tracking, and security controls with zero code changes.

📖 How the Connection String Works

1

Submit this form

Fill in your LLM provider endpoint, authentication details, and a description of what your agent will do. This creates a pending connection request that an admin will review.

2

Admin reviews & approves

An AshLab admin will verify the endpoint, review the auth model, and either approve or deny the request. You will be notified with an MCP connection string on approval.

3

You receive an MCP connection string

The connection string is a JSON object that tells your agent how to talk to the LLM via the AshLab MCP Gateway instead of calling the provider directly.

{
  "mcp_server": "https://mcp.ashlab.cloud",
  "connection_id": "conn_abc123",
  "endpoint": "https://api.openai.com/v1/chat/completions",
  "auth_method": "api_key",
  "gateway_token": "gw-eyJ…"
}
4

Route traffic through the gateway

Instead of calling your LLM provider directly, POST to the MCP bridge using your gateway token. The gateway forwards the call, handles auth, logs every request, and returns the response.

// Your agent code (example — Node.js)
const mcpConfig = JSON.parse(YOUR_MCP_CONNECTION_STRING);

const response = await fetch(
  mcpConfig.mcp_server + '/bridge/request',
  {
    method: 'POST',
    headers: {
      'Content-Type': 'application/json',
      'Authorization': `Bearer ${mcpConfig.gateway_token}`
    },
    body: JSON.stringify({
      connection_id: mcpConfig.connection_id,
      method: 'POST',
      body: { model: 'gpt-4o', messages: […] }
    })
  }
);
5

Automatic telemetry & governance

Every call routed through the gateway is automatically logged with token counts, latency, cost estimates, agent identity, and full prompt/response (if policy allows). Admins can revoke access at any time from the portal.

❓ Frequently Asked Questions

Will the gateway store my API key?
Your API key is stored encrypted at rest in the gateway's secrets vault. It is never returned in plain text after submission. The gateway uses it to call the LLM on your behalf — you never need to pass it again after approval.
Does every call go through the AshLab server?
Yes. The MCP connection string points to https://mcp.ashlab.cloud. The gateway forwards your request to the real LLM endpoint, injects the stored auth, and returns the response. This adds a small amount of latency (typically < 20 ms) but enables full observability.
How long does approval take?
Typical approval SLA is one business day. Admins are notified immediately when a new request is submitted. You can check the status at any time in the Pending Approvals tab.
Can I test the connection before my agent is in production?
Yes — once approved, your connection will appear in the Test Space LLM dropdown. You can run live queries against the gateway from there to verify the connection string works before wiring it into your production agent.

📝 New Connection Request

The full URL the gateway will forward calls to.

Leave blank → Authorization: Bearer [YOUR_KEY]

ℹ️ What happens after you submit:

  1. Your request enters the Pending Approval queue — visible to AshLab admins.
  2. Your credentials are encrypted and stored in the gateway vault; they are never exposed again.
  3. Once approved, you receive an MCP connection string (JSON) via this portal.
  4. Use the connection string in your AI agent or test it in the Test Space.
  5. All calls routed through the gateway are logged with full telemetry (tokens, latency, cost).

AshLab AI Governance Platform — questions? Contact your enterprise admin or visit the portal.