`ENETUNREACH` in Codex cloud

Open 💬 10 comments Opened Oct 5, 2025 by jakeleventhal

💡 Likely answer: A maintainer (github-actions[bot], contributor) responded on this thread — see the highlighted reply below.

What version of Codex is running?

0.44.0

Which model were you using?

gpt-5

What platform is your computer?

codex cloud

What steps can reproduce the bug?

I am trying to use the Vercel SDK with this code:

await vercel.deployments.createDeployment({
  // ...
});

an get a ENETUNREACH error from codex despite having unrestricted web access

What is the expected behavior?

Codex should be able to access the web here.

What do you see instead?

_No response_

Additional information

Found a similar issue in the community but no clear fix suggested/made.

View original on GitHub ↗

10 Comments

github-actions[bot] contributor · 9 months ago

Potential duplicates detected:

#3136
#3642

Powered by Codex Action

WillCalderwood · 9 months ago

I'm having the same problem. All internet access allowed, but test calls to AWS services are failing with ENETUNREACH.

etraut-openai contributor · 7 months ago

I think this bug is no longer relevant. If you're still seeing this with the most recent version of codex web, let us know.

cjnqt · 7 months ago

@etraut-openai
This is still happening. With all internet access allowed, test calls from node.js to public internet (for example to some AWS services) are failing with ENETUNREACH.

The setup script has internet access, for example apt-get works, but after setup the internet access is shut off even though the setting is always allowed.

Node-related?

etraut-openai contributor · 7 months ago

OK, thanks for the confirmation. Reopening issue.

rebelliard · 4 months ago

I have this error as well, when trying to have Codex Cloud interact with the Linear API:

`fetch failed` with root cause `ENETUNREACH` when trying to reach `api.linear.app`.

jakeleventhal · 4 months ago

Any way to move around this?

bselleslagh · 4 months ago

Also have this issue, for example I am unable to authenticate with WorkOS using the client credentials flows

Commands run
✅ bun run dev -- --host 0.0.0.0

❌ node -e "fetch(process.env.M2M_TOKEN_URL,{method:'POST',headers:{'content-type':'application/x-www-form-urlencoded'},body:'grant_type=client_credentials&client_id=x&client_secret=y'}).then(r=>r.text().then(t=>console.log(r.status,t.slice(0,120)))).catch(e=>{console.error(e);process.exit(1)})"

❌ curl --noproxy '*' -4 -sS -o /tmp/m2m_resp2.txt -w '%{http_code}\n' -X POST "$M2M_TOKEN_URL" -H 'Content-Type: application/x-www-form-urlencoded' --data 'grant_type=client_credentials&client_id=x&client_secret=y'

neoromantic · 3 months ago

I can't even install deps with fully open access for codex cloud agent:

On bun install postinstall script fails: error: postinstall script from "inngest-cli" exited with 1

📦 Installing [1557/1633] 📦 Installing [1604/1633] 📦 Installing [1611/1633] 🚚 lru-cache... 🚚 lru-cache... 🚚 lru-cache... 🚚 lru-cache... 🚚 lru-cache... ⚙️ inngest-cli [8/10] ⚙️ lefthook [8/10] ⚙️ supabase [8/10] ⚙️ supabase [8/10] ⚙️ supabase [9/10] FetchError: request to https://github.com/inngest/inngest/releases/download/v1.17.6/inngest_1.17.6_linux_amd64.tar.gz failed, reason: connect ENETUNREACH 140.82.112.4:443 - Local (0.0.0.0:0)

RicSala · 3 months ago

Hi,

I am also getting this issue when trying to generate basehub types (which makes a request to "https://basehub.com/api/..."

That's despiste granting internet access to all domains, all methods.

Is there workaround in the meantime?