🧐 Your opinion needed. Automations in Fibery (limitations and opportunities)

[julien.erny]

1. The ability to union lists of the same type is hands on the first on my list. But using formula in filters and sorts would a close second. A more evolved language allowing loops and variables would also be fantastic. Like a simple scripting language that would allow to open it to more users without causing too much security issues.
2. I’m using an external workflow but not sure it’s due to lack in automation. I’ve created a shortcut to send “brain dump” kind of inputs without getting out of my workflow (i.e from anywhere on my computer or my phone) to be reviewed on Fibery later (GTD-style inbox)
3. Can’t think of anything really

[jurgenappelo]

I use Make extensively. Please don’t try to build a weak copy of Make or n8n. Those tools already exist. Leave the generic automation (like integrations with other tools) to them. What’s the benefit of being able to send a message to Slack directly from Fibery? Sure, it saves me one credit in Make. Big deal. The Make scenarios are much more advanced than the Fibery rules. Let’s keep it that way.

Instead, focus on what is specific to Fibery: automations that trigger on any kind of change in the data or data structures. I’d like to trigger a rule when a field is added to a database. Or when a view of a database is changed. Or when the number of entities in a database exceeds a certain number. Etc.

And I’d like more complex flows in the rules, like if statements and switches. The workarounds that I have now to achieve conditional flow make everything quite hard to follow.

The single most important thing for me in Fibery for me is maintenance features on data structures: how is which field used and where? The flexibility in Fibery is awesome but, because of the many changes in the structures, the data models easily end up with fields, formulas and relations that aren’t being used anymore. I have no idea how I can get insight on that. I want Fibery to tell me which parts of the data structures serve no purpose anymore.

[lars]

For me, the most annoying limitations of automations (not counting JavaScript actions) are as follows:

• No conditional actions (sometimes conditions cannot be fully expressed as rule filters)
• No loops, hard to do any actions on some related items (for example, it is basically impossible to create “radio button” checkboxes on a list of items without resorting to buttons or scripting; or from query result create new linked entities in a loop)
• Also no iterating over formula results. Example use case: have a comma-separated list of e-mail-addresses in one field (from outside system, for example), link corresponding users to entity.
• For button automations: Better feedback to the user, i.e. positive or negative messages displayed, maybe even asking for input in between steps. Currently, you need to resort to scripting for user feedback (by throwing errors for example)

And sometimes, even automations are a necessary workaround because formulas are not mighty enought: I would love to dynamically query objects in a formula that have no existing relation to the entity the formula is computed on.

If I had to pick a single thing that is more important than the automation improvements, I would agree with @mkervo and choose “better workflows” (stricter state and transition handling, enforce comment/entity change on state transition, better forms with default values, etc.) – and I actually feel that this is more important than a rewrite of automations, because scripting – while tedious and annoying – does already work around the limitations listed above quite well.

[mdubakov]

@Tommy_Hedley

Polymorphic/multi-relations, I keep hoping to see these on the roadmap but I’m worried they’re gonna stay on the backburner.

In plans for next year actually, but can’t promise anything yet.

[mdubakov]

Fantastic feedback so far, folks, very helpful!

[Sev]

I have to agree with the message below.

I apologise if my answer seems off-topic, but given that the scripts in automations cover a vast number of use cases, I see improvements in the formula field (maybe even allowing JS in them) as a more impactful feature, because I have seen several formula-shortcoming workarounds in the forum that cause the use of automations instead.

As for automations, between the super-powerful scripting in Fibery, and external tools like pipedream, I cannot imagine running into serious shortcomings.

@RonMakesSystems

Nothing. If the no-code can’t do it. We can script. Anything is possible there. I think this topic is relevant for this discussion: Feedback needed: scripts in automations that shouldn’t be scripts.

[Yuri_BC]

1. What use cases you imagined with Fibery, but can’t do them because of poor automations?

   • Build Quick AI Operations Chains visually. To have easy access and build logical relations through 2D Logical Views, without having to access the Automation admin UI). Idea: Most integrated way could be Automations and Automation Actions as entities.

2. Do you compensate that with n8n, API, Make, Zapier or other external tools?

   • n8n, but for internal operatons inside Fibery workspace, n8n is overengineering; n8n is for external data operations.

3. What a single thing more important for you to have in Fibery instead of better automations? Please do not provide many, just one

• Viewer context - so Entity Views know which users are viewing the enity. I have a chrome extension of that, but should be out of the box. Massive benefits.

[Yuri_BC]

Related idea about Automations

Automation management using databases + Prompt to create script comment

Automations in Fibery I find currently ‘underdeveloped’ in that management of automations and scripts is not easy. Seen the focus on AI and workflows, this becomes even a greater need, so its likely that at some point this will be improved. In the meantime I have the following solution: Automation & Script management using databases Scripts in Fibery are dependent on the correct automation setup. This is how I manage that: ‘Automation’ database (with rich text field describing the automation…

[Yuri_BC]

Related also

Improving AI Workflow Tools in Fibery

I can’t help but continue to dream how powerful AI Agents & Workflows could level up Fibery 1,000x. In hope of some community feedback or further collaboration, I’m sharing some initial progress I’ve made while tinkering with implementing an Agentic Workflow workspace in Fibery, albeit with all of its limitations. I’ve published this public Fibery whiteboard with my current thought process and concept progress. [CleanShot 2025-04-25 at 03.00.12] [CleanShot 2025-04-25 at 02.35.57] KEY CONCEP…

[myg_ge]

What would be really useful is also having shared functions across formula fields. That would allow to overcome current UI-level limitations of presenting data while not requiring creating redundant formula-level piece of functionality over and over like this one:

ToText(Year(Deadline)) +
  If(
    Month(Deadline) < 10,
    "0" + ToText(Month(Deadline)),
    ToText(Month(Deadline))
  ) +
  If(Day(Deadline) < 10, "0" + ToText(Day(Deadline)), ToText(Day(Deadline)))

just to show date with leading zeros.

[Sev]

@myg_ge

What would be really useful is also having shared functions across formula fields.

Agreed. To avoid getting further off-topic, I have found a handful of somewhat-related topics, but the closest one is this: Variables in formulas. I am moving my response there .

[Sppam_Kontrol]

we have issue with the amount of automation allowed per plan. it is too limiting.

1. Imagine, you just want to move one doc to another in a self-relation. It is already counted as 1 automation.
2. So, there is no notification truly embedded in the platform. Just creating a notification already eats up credits using the automation as well.
3. there is also a lack of support for automation with the docs. we would like to do automation on exporting docs and/or entries to a pdf where the pdf structure/page properties are also configurable. ZUsing JS alone is not practical as we have to reconfigure all the codes again even with just a small changes from the sources.
4. Request for a scheduler management for automation. these includes remaining credit, predictive utilization, manage schedules, create cascading automations (like after 1 automation is done, it can perform the next one) this is to create a better management of chunks of automation in a intended automation pipeline.

[Chr1sG]

@Sppam_Kontrol

Imagine, you just want to move one doc to another in a self-relation. It is already counted as 1 automation.

Not sure what you mean here. I guess you’re not talking about documents given that you’re referring to self-relations. Why would changing how entities are linked via a self-relation count as an automation?

[Sppam_Kontrol]

Similar to the template for the the.fibery.io, under the DB’s Rules, there is propagate section to children and section change to children. those are 2 rules. And everytime we just move the subdoc to a different doc category. it is already treated as an activity.

Our opinion in case is just a example where similar small items such as these ones eat alot of automation credits especially if there are a good amount of there applied across similar dbs.

[helloitse]

@mdubakov

What a single thing more important for you to have in Fibery instead of better automations?

Improved complexity with collections in formulas (or computed lookups). I’d like to directly calculate filtered lookups without needing helper fields/calculations, to control the sort order, and specify the number of results. To me, this is preferrable to the planned collection view update to lookup fields

Allow Formulas to return a collection of just a specific field from Filter results

I have a Collection of “Roles”, where each Role is related to a User. I want to create a Formula field that returns the collection of all the Users linked to the Roles in the collection. But it’s not currently possible to return a collection of a particular field from the Filter results: [image] Currently we can only do this kind of operation for a single result, not for all items in a collection: [image] IIUC this can currently only be done by using a pair of Rules (Role Linked and Role…

[Sppam_Kontrol]

Also, I think that automation credits should be increase. Image just for the hiring team, even when sending a rejection email that takes a 1 rule in automation. Imagine for just 1 role, we have like 150-200 applicants, that already eat 1 user seat automation allocation in just a few days.
It would be nice to have like a separate automation usage for like internal to the process of fibery vs external to fibery. there are automation internally that are just common features for some apps like notifications and we have to put automation just to have it.

[Fernando_Fontoura]

Alternatively or concurrently, extra credits to purchase would also help

[Mircea_Braescu]

Both cases
an example for the same DB is exactly the one I mentioned.
I might want a “report issue form” where the priority text is
“High - critical features are not working or data loss imminent””
while for the devs on the receiving end, I want a compact view where they only see
“High” in the Priority column.

[Chr1sG]

@Mircea_Braescu

I might want a “report issue form” where the priority text is
“High - critical features are not working or data loss imminent””
while for the devs on the receiving end, I want a compact view where they only see
“High” in the Priority column.

You could have a priority select field, where the Name is High, and then add a Text field to the select options, where the value for High would be “Critical features are not working or data loss imminent”.
Then the issue is that form view does not currently allow for showing additional fields in the dropdown, right? Assuming that were fixed*, is there any reason for having two identical select fields on the main db?

*Related topics:

Display fields in "link or create" window

I have many things in a single database that are all named the same. For example, in my “Iteration” database I already have over 200 records called “Iteration 1” and there will ultimately be thousands more named the same way in the future. I wish that when I clicked on “Link or create”, I would be able to view some fields (like “Client” or “Due Date” or “Team” for example) that are on the entity I’m trying to link to. Currently the only fields that are shown are “Name” and “ID” and “Database”…

Open larger view on relation field dropdown

[Screenshot 2025-02-21 at 11.06.24] Problem: When choosing an entity to relate to, you might have many to choose from and not sure which on is that you are looking for. Either because of duplicate names, or just because you need to explore a little before choosing the right one. Right now I need to go a seperate view, explore / filter / search, find the one i need, go back to the original entity and link it or link it from the target entity, then go back to the original to make the rest of the…