See every outbound call
your agent makes.
Agent egress monitoring shows the outbound network connections an AI agent makes. The AiDren worker-agent npm package runs inside your agent's process, learns normal destinations in training mode, then flags any connection to a destination off your allowlist on your Events page. It reports connections and never blocks them.
How AI agent egress monitoring works
A watcher inside your process.
Install one npm package and import it at startup. It observes outbound connections and reports them; it is not a hard firewall, so it cannot take your agent down.
-
1
Add the package
npm install @aidren/worker-agent, thenimport '@aidren/worker-agent/init'as early as possible in your app. Set your proxy key in the environment. -
2
Watch destinations appear
Observed destinations show up on your Egress page. It starts in training mode while it learns what normal traffic looks like.
-
3
Lock the list
Once the list looks right, lock an allowlist. Any connection to a destination not on it raises an
unexpected_egress:<host>:<port>event on your Events page — flagged, never blocked.
Visibility, not enforcement
It reports what your agent connects to. It never severs a connection.
- Every outbound destination your agent’s process opens, by host and port.
unexpected_egress:<host>:<port>events for anything off your allowlist.- Training mode to learn normal before you lock a list.
- No blocking — a real connection is never cut; you get the alert, not an outage.
- Off your chat allowance — egress reports do not count as checked requests.
Related: OSS / third-party agent security · Prompt-injection blocking · How AiDren compares →
Flagged destinations land on your Events page.
A real screen from a live AiDren account.
How much latency does AiDren add?
The proxy layer itself adds about 5 ms. On top of that, requests that need screening get one classification call — usually 200–500 ms, which runs before your request reaches OpenAI, Anthropic, or Mistral, so it overlaps nothing. Short, clearly-benign messages skip the classification call entirely. Model-file scans and egress reports add nothing to your chat traffic.
How many requests do I get?
A "checked request" is one AiDren screens on its way to your model. Starter covers 100,000 a month, Growth 500,000, Scale 2,000,000. It's a fair-use guide, not a hard limit: if you go over, we email you about a 10,000-request top-up (£19) or moving up a tier, we don't cut you off mid-month. The count resets on the 1st. For very high volume, email us for a custom plan. Model-file scans and agent egress reports don't count.
Do you see my data?
Your requests pass through AiDren to make a block/allow decision and are never stored beyond the decision log you see in your own dashboard. Your API keys are encrypted at rest and never logged in plain text.
What happens after the trial?
Your 14-day trial needs no card up front and covers 25,000 checked requests. When it ends, upgrade to keep your proxy running — if you don't, AiDren pauses your traffic rather than silently letting it through unprotected.
Is egress monitoring the same as an egress firewall?
No. A firewall blocks; the AiDren worker-agent only reports. That is deliberate: it will never cut a legitimate connection. If you need enforcement, use a network-level allowlist alongside it, and use AiDren's flags to learn what that allowlist should contain.
Which languages does the worker-agent support?
It is an npm package (@aidren/worker-agent) for Node.js apps. The API reference lists no other runtime.
Know where your agent
is actually connecting.
Get the full stack free for 14 days. No card, no sales call, no SDK rewrite.
Start protecting requests