This is a follow up on my earlier post about Selemium…
So I decided to take a look at Axiom. It doesn’t integrate with Airtable direclty, it can work via Make or Zapier but they integrate with Google Sheets nicely.
Axiom is an interesting potential solution tool to scrape data but also to take the Airtable data to fill in 3rd party forms, and this second option is what I was looking at. For scraping there are many other apps already including Bardeen, but there isn’t really many options when it comes to form filling using Airtable data.
Axiom configuration is surprisingly robust and easy with probably every typical scenario covered. However there are several issues with it that make more expensive than hiring a human being on taskrabbit or upwork to enter the forms manually. The desktop application uses Chromium runtime to allow interaction with local files and forms behind login screens. Overall the concept works, there were minor issues when the extension would not see the Chromium runtime is running but that can be resolved by leaving the small notification about the runtime running open all the time and not closing it as it suggests). The interaction with password procted forms (sites behind login) is problematic, while Chromium runtime keeps login sessions active as you would expect, there are frequent cookie noticed when the automation is re-run. I guess that can be handled (maybe) by installing extensions to prevent cookie popups or by adding steps into Axiom automation, but this just adds to the expensive runtime (more about it below). If there are forms that require a mixed input, e.g. manual input while the automation is running or even paused, this is not possible to do using the cloud solution so requires the desktop app. The desktop app unfortunately takes all the local resouces. So if I have something else running on the machine that is CPU intensive, the automation speed in Axiom is taking a hit too. For instance, typing in text into the form field slows down if I have the machine performing other tasks. While new hardware shouldn’t see much impact, using older machine and running other tasks on the machine, would bring Axiom automation to a crawl. I am not so sure if this doesn’t have an impact on the run-time consumed, because the speed of the automation is measured from the beginning the automation starts till it completes. Axiom allows adding scripts to steps, but that requires programming knowledge. I wish there were more configuration options baked in for checking conditions such as if the automation should run or wait depending on a condition met (e.g. text on the page). Some forms require manual interaction e.g. if a form is a multi-step form and requires many different selection subcategories and these subcategories change and they depend on category selected etc… So I need manually select that first and I don’t have any data to supply to Axiom to automate it for me because I don’t know myself what I need to click ahead of time. This requires me to manually interact with the form first, so I have to add a wait / delay step. The only way to do that is to use the desktop app because the cloud Axiom app does not allow me to interact with the form in any way. Adding wait time however, appears to add to the runtime of the automation, so you end up ranking up runtime because wait step is involved with every run of the automation. And this brings me to the overall conclusion.
Axiom pricing model is ridiulous.
They are charging for the runtime regardless whether the automation runs in the cloud, or on the local machine in the Chromium browser. The automation runtime includes wait time if I need interacting with the form. The runtime depends on the speed of the machine, the slower machine, the more runtime required, the more cost is going incurr.
In the age of Zapier and Make it is easy to get used to freemium models and relatively inexpensive cloud automation and the issue with that was mentioned in other posts earlier by @bfrench for example. However in the case of Axiom we are not talking about an automation platform that would be handling some volumes of data or traffic or ping other sites for changes etc. All in Axiom is basically interaction with text and forms and local application uses my own CPU and bandwidth.
If it the cloud app was at least capable to interacting with web-forms and populate attachments such as pictures I have in Airtable (or the exported expiring URLs to these in the CSV / Google Sheet document), the argument could be the cost of runtime includes loading the pictures into the online form using the Axiom infrastructure. But why would I end up paying for the runtime of the Axiom application when all the heavy lifting is done on my local desktop using the Chromium runtime and my bandwidth.
They charge by minutes/hours the automations running. It’s an unusual concept to charge by time. But the way the application counts the runtime is weird. A multistep form with text entry and loading pictures can be completed by a skilled typist in about 30 seconds at max. Because the form requires my interaction such as selecting through categories/subcategories, multi-step forms that are howering over the page and entering long text into the field boxes and then it will do loading images which also counts to runtime. Once the form is completed, the autiomation will close the Chromium runtime window.
That is not what I want. I need to review it before it is submitted, so I can’t let it handle by Axiom. There may be a way to just stop the automation but I didn’t figure out how, so I ended up adding wait time at the end of the automation, review the imput and then let it submit the data in the form.
This all requires me to add wait time into the automation so I end up with a workflow that runs 2-3 minutes because all the steps needed above. I am sure there is a better way to configure the form, but I am not an expert, nor I want to become one. When I open Make or Zapier, I am greeted with something that I have control over even if I have no experience with setting up an automation flow before. They are a clear productivity gain. On the other hand, due to the nature of it, Axiom automation requires me to be present and interact with the form but that is not an issue. The issue is the Axiom price model.
Overall, the product is well done, but for me it is more expensive to pay for Axiom than to hire somebody to fill forms. Other issues are cosmetic, such as if I need to interact wit hthe form, the Axiom panel is too big so it is over the whole page and I need to constantly close it and open it. Not a productivity gain.
In the end, I was excited about this solution and thought this was a real gem, exactly what I was looking for. But the pricing model makes this the practical use a non-starter. For a larger business, paying unlimited subscription to this is probably a no-braner, but for a smaller user, inspite of the simplicity of the interface, this is a non-starter.
I hope there will be other tools like this in the future because this is an area where I find it is a niche… (SCRATCH that… just found out Browserflow also supports form filling. whahaa, I’ll look at that )… in the browser and interacts wit the page, so hense the concept of charging for the runtime. That is examply why I think the interaction should be based on the number of automations completed and not based on runtime.
Overall - high expectations initially - but pass. No way. Continue with outsourcing work. At least I am doing a good deed to somebody who feeds a family.
Just for comparison - Axiom mentions on the price website:
Runtime: You will be able to scrape approximately 300 pages per hour or populate approximately 180 forms per hour. Runtime can vary depending on the content you wish to scrape or the amount of steps it takes to complete a bot.
That may… may … be the case for scraping websites, if your machine is latest specs, but no way to interact with a form as described earlier. In my calculations, I would be able to fill in about 20-30 forms for the 1 hours. If I a person interacts with the form, would be able to complete the same form in about 30-40 seconds (I know it because I can do that myself). So the cost for me hiring somebody would be about 2/3 of the cost I pay for Axiom.
Axiom is overpriced. That is the right word for it, given that all heavy lifting is done on my own machine and not in the cloud. I understand businesses make their own pricing decisions and many price by the value they provide. I do not think Axiom provides more value than hiring a person.
Here is pricing for Axiom - note the pricing is for total runtime per month, at the low level you get just 5 hours of extension running time in your own browser on your computer and that will cost you 15 dollar. There is no difference running Axiom in the cloud or your or browser, still paying the same money.
Here is example of how Browserflow priced their solution which makes more sense, the Time per run is limited for an individual run when running on the local machine in the web-browser, there is no total number of runtime you can have on the local machine. The time is limited to the number of minutes if running in the cloud.
So next I am going to text Browserflow. To my surprise, Browserflow now allows interacting with forms and filling in!
That is new feature, I don’t remember seeing it there when I look at them a few months ago.
Let’s take a look at that.