Streamlining Image Moderation API Uploads in Node.js: From Native Fetch to Form‑Data and DNS Fixes ⚙

In this post, we’ll walk through our journey integrating the Sightengine image moderation API into a Node.js application. Along the way, we encountered and solved several challenges:

1. Native fetch + form-data incompatibility
2. DNS resolution failures in Node.js
3. Blob vs. stream limitations
4. Final streaming solution with form‑data + node‑fetch

1. The Initial Approach: Native fetch + form-data

We began by trying to mirror our working curl -F snippet in pure Node.js using the built‑in fetch (Undici) and the popular form-data package:

import FormData from 'form-data';
import fs from 'fs';
const data = new FormData();
data.append('media', fs.createReadStream('./image2.webp'));
…  
await fetch('https://api.sightengine.com/1.0/check.json', {
  method: 'POST',
  headers: data.getHeaders(),
  body: data
});

However, this led to a ConnectTimeoutError and eventually a DNS ENOTFOUND—even though curl worked perfectly.

2. Hunting Down DNS Failures

A quick nslookup api.sightengine.com returned:

;; Got SERVFAIL reply from 127.0.0.53
server can’t find api.sightengine.com: SERVFAIL

Our system was using systemd-resolved on 127.0.0.53, which silently failed. The solution was to bypass this resolver and point to public DNS:

# /etc/systemd/resolved.conf
[Resolve]
DNS=8.8.8.8 1.1.1.1
FallbackDNS=1.1.1.1

After sudo systemctl restart systemd-resolved, both nslookup and our Node.js code could resolve the API host.

3. Native FormData and Blob vs. Streams

Next, wanting to stick with pure native APIs, we tried using:

import { Blob } from 'buffer';
const stream = fs.createReadStream('image.jpg');
const blob = new Blob([stream], { type: 'image/jpeg' });
formData.append('media', blob, 'image.jpg');

But Undici’s FormData.append still only accepts string, Buffer, or real Blob contents—not raw Node streams—so the API returned a “could not read image” error.

4. The Final, Battle‑Tested Solution

We ultimately opted for the most reliable approach in Node.js land:

1. Install

   npm install form-data node-fetch

1. Code

   import fs from 'fs';
   import fetch from 'node-fetch';
   import FormData from 'form-data';

   const form = new FormData();
   form.append('media', fs.createReadStream('./image.jpg'));
   form.append('models', 'nudity-2.1,weapon,offensive-2.0,gore-2.0,self-harm');
   form.append('api_user', 'your-api-user');
   form.append('api_secret', 'your-api-secret');

   const res = await fetch('https://api.sightengine.com/1.0/check.json', {
     method: 'POST',
     body: form,
     headers: form.getHeaders()
   });
   const result = await res.json();
   console.log(result);

This mirrors curl -F exactly, streams the file (no big memory hit), and produces valid multipart requests.

Conclusion

Integrating file uploads in Node.js can feel straightforward—until you hit native vs. library quirks and platform DNS oddities. By:

1. Fixing your DNS resolver
2. Recognizing Undici’s FormData limitations
3. Falling back to form-data + node-fetch

you’ll have a rock‑solid, streaming multipart setup that behaves just like curl -F.