Skip to content
Fixedmark
Use case · Automation

Static IP for AI Agents and n8n Automations

AI agents and automation tools call APIs from cloud infrastructure with changing IPs, so allowlisted APIs reject them. Route those calls through a Fixedmark proxy URL, using the n8n HTTP Request node's proxy option or a proxied HTTP client in your agent's tools, and the API sees one dedicated IP pair.

Why automations hit IP allowlists

Automation platforms and agent frameworks run on shared cloud infrastructure. n8n Cloud, Make, and hosted agent platforms send requests from address pools that you do not control and that change over time. When a workflow or agent calls an internal API, a customer's ERP, or a vendor that allowlists IPs, the request is dropped.

Self-hosted n8n or an agent on Railway or Render has the same issue in a smaller form: the container's IP changes when it redeploys or the platform moves it.

n8n, Make, and Lovable

In n8n, the HTTP Request node has a Proxy option under Options. Paste your Fixedmark proxy URL there, and only that node's requests go through your IPs. See the n8n guide for details.

If a no-code tool has no proxy field, put a small function between the tool and the API. The tool calls your function, and the function calls the allowlisted API through Fixedmark. Lovable apps run their backend logic in Supabase Edge Functions, which can use a proxy through Deno.createHttpClient, as shown in the Supabase guide.

AI agent tool calls

An agent's tools are ordinary HTTP calls made by your code. Give the tools that hit allowlisted systems their own proxied client, and leave model API calls on the direct path. That keeps proxy bandwidth low and makes the logs readable.

Agents choose their own requests, so limit what they can reach. A destination allowlist on the Fixedmark token, such as erp.customer.example:443, means a prompt-injected agent cannot use your static IP to reach other hosts. Connection logs record every destination the agent contacted.

Give an agent or workflow a static IP

Use the same proxy URL for every workflow and agent that needs the allowlisted IP.

Fixedmark is in early access. Proxy URLs are issued at launch. Join early access to get yours when your region opens.

  1. 1
    Create an endpointChoose the region closest to the API your workflows call.
  2. 2
    Allowlist both IPs at the destinationSend both addresses of your pair to the API owner or add them to the vendor's allowlist.
  3. 3
    Add the proxy URLIn n8n, set it on the HTTP Request node's Proxy option. In code, store it as FIXEDMARK_PROXY_URL.
  4. 4
    Proxy only the tools that need itCreate a proxied HTTP client for allowlisted tools. Keep model and public API calls direct.
  5. 5
    Restrict destinationsAdd a destination allowlist so the token can reach only the hosts your agent is meant to call.
HTTP Request node
  Method: POST
  URL:    https://api.partner.example/v1/orders
  Options
    Add option > Proxy
      Proxy: <your Fixedmark proxy URL from the dashboard>

Frequently asked questions

How do I give an n8n workflow a static IP?

Open the HTTP Request node, add the Proxy option under Options, and paste your proxy URL. Requests from that node then leave from your static IP pair.

Can I use a static IP with Make or Zapier?

Only where the tool exposes a proxy setting. Otherwise call a small function you control, and have that function call the allowlisted API through Fixedmark.

Should my LLM API calls go through the proxy?

Usually not. Model providers do not need your static IP, and proxying them uses bandwidth. Proxy only the tools that call allowlisted systems.

How do I stop an agent from misusing the static IP?

Set a destination allowlist on the token so it can reach only approved hosts and ports, and review the connection logs.

Make it fixed.

Fixedmark is in early access. Join the list to get dedicated static IPs when your region opens.