Need to scrape data from websites like LinkedIn, Amazon, or real estate platforms? I build custom Apify Actors that get exactly what you need. DM me if you need help with a web automation project!

Doesn't this type of messages trigger the exact opposite of the intended effect? 😂 stop spamming please

Ohh ok

No spamming

thanks for reporting.

Hello 👋

@Temitope just advanced to level 1! Thanks for your contributions! 🎉

Do we have the feature to add Global environment variable so that we can set it up at one place and make it utilize by all the actors. For now we have to set it up on all the actor individually. though its a common variable.

There isn't anything like that but good idea!

It would be great if we had an option to give more trial time for users that opened an issue.

It would be very helpful to have default visibility into private actor runs when sharing it (run permission) with someone, or at least the option to enable this.

Can we please let earning actors talk to support without having to go trough the AI chat? If I can't change my creditcard details I don't want to chat to an AI

Stand by actor monetization

Why start docker image each run it would be best if there is server up 24h to receive api calls

This is in progress but the final form will take a few months

I’d really like to be able to change the default BUILD_DEFAULT_MEMORY_MBYTES value on Apify. Ideally, there should be both a minimum and maximum memory setting available when building the Docker image just like we can define memory limits for actor runs. My goal is to optimize for the lowest cost possible. For example, I already compile my Rust code locally and upload only the final binary, along with a minimal image that simply runs the executable. With this setup, my build now costs only $0.005, which is already a big improvement. However, I noticed that it still defaults to 4096 MB of memory, even though my Docker image is minimal and the compilation is already done. That amount of RAM seems excessive for this kind of lightweight build. The first time I built the image when I was compiling the Rust code inside the Dockerfile the cost was much higher, around $0.10. Now it’s more optimized, but I believe I could reduce the cost even further if I were allowed to set the memory to something like 128 MB. Having options like buildMinMemoryMbytes and buildMaxMemoryMbytes, or even better, the ability to override the default value with something like buildDefaultMemoryMbytes, would give developers more control and flexibility especially when builds are frequent and optimizing resource usage is critical.

I understand your concern and I was afraid of useless ram overusage too but fact is actor will eat only that amount of memory it actually needs, so if you have leightweight actor it will use little ram regardless of default value. At least this is how I understand it, the RAM isn't somehow reserved and charged in that default value amount, only real usage is

I believe the way it works is that no matter how much RAM the build actually uses, you will always be charged assuming 4096 MB, which is the default value. Here they mention that the build memory size is 4096 MB: https://docs.apify.com/platform/limits And in their internal code on GitHub, the variable also appears hardcoded like this:

// The actualy used limit is taken from private package @apify-packages/consts

BUILD_DEFAULT_MEMORY_MBYTES: 4096,

https://github.com/apify/apify-shared-js/blob/4a7608fbdd51dbc02661f4be03a4a1d3bb91cd2b/packages/consts/src/consts.ts#L295 Here they explain how CU consumption is calculated: https://docs.apify.com/platform/actors/running/usage-and-resources#memory “A compute unit (CU) is the unit of measurement for the resources consumed by Actor runs and builds. You are charged for using Actors based on CU consumption. For example, running an Actor with1024MB of allocated memory for 1 hour will consume 1 CU. The cost of this CU depends on your subscription plan.“ https://cdn.discordapp.com/attachments/1075487274424352888/1397245798076387449/image.png?ex=6883a90d&is=6882578d&hm=7fbfaf629ef67ffa5f45d8adfc33febacba6e39f8f5514b0ca74b588ec8162ae&

So:

1 CU = 1024 MB * 1 hour

So the formula would be:

CU = ( MEMORY (MB) / 1024 ) * DURATION (HOURS)

If we use the default value, which is 4096 MB, then for example with the 13 seconds of my build image:

Memory in CU = 4096 MB / 1024 = 4 CU/hour 

Time in hours = 13 seconds / 3600 = 0.0036111111 hours 

Total CU = 4 × 0.0036111111 = 0.0144444444 CU 

Total cost = 0.0144444444 × $0.40 = $0.0057777777

Rounded, it would be $0.006

I sometimes had struggled with not having enough build memory but it was always a niche problem so we are not sure if it is worth options. Are you building so often that these costs become real?

As a developer who primarily uses Rust and Golang, I prefer to compile my binaries locally, ensuring that the architecture is compatible with Apify’s infrastructure. Then I create a minimal Docker image that simply executes the binary, and I upload both the binary and the image to Apify. Thanks to this strategy, the actor doesn’t require any build time on Apify’s servers, and its execution time is very short. In fact, I’m fairly confident it could even run with just 128 MB of memory, since all heavy tasks have already been handled locally. Reducing memory requirements during the build could also help lower the base price of actors, as build costs beyond the free tier are covered by the developer. This would reduce friction for new users to try out the actors, and at the same time, it would ease the load on Apify’s servers. Overall, it would be a win-win improvement for everyone involved: users, developers, and the platform itself. That said, I understand there are likely more urgent tasks on Apify’s roadmap, so I don’t consider this a priority but I wanted to share it as a future suggestion.

Hello Apify Team, I am using the Instagram Scraper Actor (maintained by Apify) to scrape public Instagram reel URLs. However, I consistently encounter the following error in the dataset: "error": "restricted_page", "errorDescription": "Restricted access, only partial data available"

Hello Apify team, this is what I have felt regarding notification. It will be so good to add TAGs or some labels on the notification message we receive to differentiate if its a support notification or price change or invoice. For now its hard to skim it, have to read , open it and check. Labels if added will be so easier.

Is it possible to run Docker on 64 MB? https://cdn.discordapp.com/attachments/1075487274424352888/1399623422014918758/image.png?ex=6889ac63&is=68885ae3&hm=54023a8178371cc1fa701b141b9a3ffe990b0e97e86e46bc7cc3e9b393d02198&

Not now, I think Node.js might start to struggle there. Do you use other languages?

I use golang, 128 mb is too much for most of my actors 😅 https://cdn.discordapp.com/attachments/1075487274424352888/1399868861435875398/image.png?ex=688a90f8&is=68893f78&hm=1b2c994deb12db2d300006a976c1647b5f7785e64e3c643015bdf9bc5daa40f7&

Pay-Per-Event Pricing Display Issue - Making Actors Appear 100x More Expensive I've noticed a serious issue with how pay-per-event pricing is displayed that's making actors appear extremely expensive when they're actually very affordable. The current display shows costs like "Actor startup fee: $50.00 / 1,000" when the actual cost is just $0.05 per run. When users see that "$50.00" number, they immediately think the actor is outrageously expensive and won't even try it. In reality, they're only paying 5 cents for a one-time startup charge. I'm experiencing this firsthand with my actor. The pricing display makes it look prohibitively expensive, so no one wants to use it. But when you break down the actual costs, it's incredibly reasonable. For example, what shows as "Actor startup fee: $50.00 / 1,000" is actually just $0.05 per run for the one-time initialization charge that covers browser setup, and system resource allocation. The memory usage pricing is equally confusing. It displays as "Memory usage (30-sec intervals): $3.30 / 1,000" which makes users think they're paying $3.30 per 1,000 results. But it's actually $3.30 per 1,000 thirty-second time intervals. So a typical 2-minute run with 4 intervals only costs about $0.013 for memory usage. The math works out to something like this: a typical run costs around $0.05 for startup plus maybe $0.01 for memory, totaling about $0.06. But users see those big numbers in the pricing display and assume it's going to cost them $50 or more. This pricing display issue is probably affecting a lot of pay-per-event actor developers. We're creating affordable, efficient actors but the pricing display is scaring users away before they even understand what the real costs are. The whole point of pay-per-event was to make pricing more transparent and fair, but the current display is doing the opposite

Hey we had what you are suggesting previously but we got feedback from both developers and users that it is very hard to read and not industry standard. Developers were ending up setting wrong pricing, sometimes an extra zero or sometimes a zero less. Users were complaining that its not industry standards to tell pricing this way, everyone tell price per 1000 events.

Hey guys I already asked in other channel but, I have a question about apify google maps scraper, is it possible to alter it in any way so that on a new run of the actor it doesnt output/scrape duplicate results of past runs?