Instagram Scraping: Apify vs a Custom bro scraper
We compared Apify's Instagram Scraper against our open-source solution to evaluate cost, speed and data quality.
TL;DR: We tested our open-source Instagram scraper running on bro against Apify.
On a dataset of 1,200 comments, our scraper was 32x cheaper. For deep comment scraping (15,000+ records), it was up to 75x cheaper. Apify is faster out of the box but running our scraper in parallel across multiple browsers sessions cuts collection time by 30% for essentially the same price.
Does scraping Instagram at scale have to be expensive? Apify is the go-to solution for many but the costs scale up fast.
Armed with bro's stealth cloud browsers, we built an open-source alternative and put it to the test.
We ran a head-to-head comparison against Apify to measure cost, speed and data quality - we matched their data but for a fraction of the price.
How we set up the test
To keep things fair, we gave both tools the exact same task:
- Target: 12 Instagram post URLs.
- Goal: Extract 100 parent comments per post (1,200 total records), excluding replies.
- Setup: We used one authenticated bro browser session, routing traffic through our lite proxies in Germany.
- Apify setup: We ran Apify actor twice to ensure that ensure comments match between runs. Actor config had default parameters.
Cost and Time: 32x Cheaper
Here's how the runs compared for extracting 1,200 comments:
| Run | Records | Run time | Total cost | Cost per 1k records |
|---|---|---|---|---|
| Apify (Run 1) | 1,200 | 77 seconds | $2.76 | $2.30 |
| bro scraper | 1,200 | 12 min 6 s | $0.086 | $0.072 |
| Apify (Run 2) | 1,200 | 70 seconds | $2.76 | $2.30 |
Our exact recorded bill for the run was $0.086, which is roughly $0.072 per 1,000 parents (including browser time and proxy traffic). Apify charged $2.76, meaning bro is roughly 32 times cheaper for this batch.
Apify returned the dataset significantly faster, finishing in just over a minute compared to our 12 minutes. However, we made our pipeline slower to reduce a risk of getting blocked.
Did the data match?
Yes. Our output shared 98.50% to 98.75% of the exact same comment IDs with Apify's exports.
Why wasn't it 100%? Instagram's feed is constantly shifting. Even Apify's own two runs (run just minutes apart) didn't match perfectly.
A 98.5%+ overlap is standard for live social media scraping and confirms we were extracting the same data.
Scaling Up: Deep Collection and Parallel Scraping
Small batches are great for testing but what happens when you need a volume? We tested deep collection runs extracting over 15,000 comments and replies to see how the costs changed.
We also tested running our scraper in parallel across two bro sessions. Each browser used its own proxy IP and separate session cookies.
| Deep Collection (15,800+ records) | Run time | Usage charge | Savings vs Apify |
|---|---|---|---|
| Apify (Estimated at $2.30/1k) | - | ~$36.43 | - |
| bro (1 session, seq order) | 3 hr 17 min | $0.48 | 75x cheaper |
| bro (2 parallel sessions) | 1 hr 27 min | $0.59 | 61x cheaper |
For deep collection at this scale, bro is practically free compared to managed services.
Beyond comments
We also tested other data types like profiles, reels and hashtags. We matched 100% of the IDs for profiles, reels and tagged posts.
Here is a quick breakdown of how bro's usage cost compared to Apify for other extractions:
- 30 Profile Posts: bro is 1.57x cheaper.
- 30 Reels: bro is 1.52x cheaper.
- 30 Tagged Posts: bro is 1.29x cheaper.
- 30 Hashtag Posts: bro is 2.73x cheaper.
Apify was actually cheaper for ultra-small extractions (like fetching specific profile details or grabbing exactly 2-3 search results for users, hashtags and places) because bro's browser startup cost eats up the tiny savings on the actual page load. But for anything at scale, bro takes the lead.
What provider to choose?
Use Apify if: You want a dataset immediately, you don't mind paying an extra premium or you don't want to maintain Instagram accounts.
Use bro if: You want full control over your extraction logic, need to run parallel sessions without getting rate-limited and want to slash your bills significantly.
With bro, you pay for raw browser runtime and bandwidth, not per extracted records.
Try it on your own data
Want to see for yourself? You can find the source code of the scraper in the git repo.
See our recent blogpost to learn more of how the scraper works.
You can configure it for your own targets, set up multiple accounts for parallel extraction and run it directly on bro's cloud browsers:
Integrate to your project with just one prompt
Focus on the data, not the infrastructure.
Copy the prompt and paste it to your AI agent.