🐴 Fibery Product Report: May 2025

Previous report: 🐢 Fibery Product Report: April 2025

Features are in bold.

We’ve made four releases in May. Total Highlights Score: 400 (for comparison, it was 440 in April).

Polina recorded a webinar, watch it to see all changes in action:

:farmer: After playing with the initial Workspace Builder for Fibery I feel more positive about its huge potential. However, the scope is also huge, we need to make this flow work with data import, full workspace creation and modification (which means context for prompt is important), integration with external tools should be way more easier (MCP?, not sure). Anyway, it seems we will need more people in this theme soon, but we also need to define the path and scope clearly first.

Current and nearest future themes (with assigned devs):

Import

Data

In late April, we started to collect data on how people use import and integrations. For example, here are the most popular connectors for import specifically over the last 30 days:

This chart doesn’t take the deprecated CSV 1.0 into account so CSV is, actually, even more dominant. Integrations converted to native DBs later are not counted either.

Unfortunately, entity syncs data got corrupted from Apr 27 to Jun 3, so we’ll have to wait a bit longer for other insights.

CSV

After two years of CSV 1.0 and CSV 2.0 beta™ co-existence, we’ve finally merged the two: Merge CSV and CSV 2.0 import, so no more confused customers asking us how to choose between 1.0 and 2.0.

We’ve picked the import based on our integrations framework as a way to go. As a side effect, we’ve polished the the framework quite a bit:

  • Prepare Field mapping UI for relations
  • Support Phone Field in integrations & import
  • Support Location Field in integrations & import

The next step would be to gather and address things people are missing from the deprecated CSV 1.0:

  • Import self-related entities via a single CSV file
  • (certainly more to come)

Discoverability

We’ve made it easier to discover import when creating a new Database:

Next, we’ll apply the same pattern in other places:

  • Suggest import or sync when adding a new Space
  • Replace “Import data” with a nested “Import from…” menu
  • Update headers for Database screens

Unfortunately, Ihar has to tend to Table View bugs so the progress has been very slow so far.

We are also in the process of democratizing import: Allow non-Admins to import data into existing Database.

Notion

We finally fetch Database Pages content from Notion in our connector, making the migration smoother (and, thus, more likely):

We will gather feedback and polish the rough edges in June as well as Sync Notion standalone Pages to Fibery Database.

With Notion being our #1 competitor and the tool people most often transition from, our connector should be top-notch.

Unique values

We’ve released the option to require values of a Field to be unique:

While this doesn’t directly relate to import, it frequently overlaps in real-life scenarios. Also, we are looking to re-use the mechanisms for Update existing entities via CSV import.

The next step is to help architects to Find and merge existing duplicates when enabling unique toggle. We won’t introduce any special UI (too much effort, too little re-use) and try to get by with a neat Table View for manual merging and a set of rules for automatic merging.

Permissions

Space → Database

Andrew is still doing his under-the-hood part that has proved to be much more effort than we initially expected. The end is near, we hope.

Next: Allow not to extend access from Space to its Databases

Manage User’s and User Group’s access

We are about to release a centralized access management for Users and Groups:

This would help:

  • troubleshooting [the lack of] access;
  • providing access to a newly created group or newly invited user;
  • reviewing permissions when someone changes roles within their company.

At first, only direct access will be shown but later we’ll move on to Introduce access audit for Users and User Groups.

Sidebar

We’ve started Home Page in Fibery. First release will be static page that will be useful for newcomers, then we will add more widgets.

  • My Work Widget on Home Page
  • Recent Items widget
  • Embed Views into Home Page

We hope to complete this theme in June.

Gantt View

All features for Gantt View were implemented (vertical DnD, units, creation of new entities, etc), now we are polishing and fixing bugs. Gantt View should be released as experimental in 1-2 weeks and we are going to complete it in June.

Views enhancements

After Gantt View we will spend few weeks on some views improvements:

  • Specify which Entity View to open based on some field value
  • Board View Improvements

AI

Q&A Smart Agent

We’ve released Smart Agent for Q&A in May 22 and have around 100 questions per day from all users in all accounts. There are 30 accounts with 10+ questions during this timeframe. I’d say usage is not significant so far. However, for Bots we have ~7 answers per day, so Q&A Agent is 10x more popular.

Common themes:

  • People try to do many things via agent, like ask user guide questions, change/add entities, change schema, create automation rule, etc. From my estimate it is about 30% questions about user guide, 30% questions about changes and 40% are real queries. It shows that Agent should work for guide, and for creating things as well.
  • Common feedback is that Smart Agent is quite slow, but quite good. Overall, not much feedback was provided.
  • Repeatability of prompts is not great, it is quite hard to automate prompt and expect consistent result for complex things, like weekly comments report. :aleh-seriaha: has an idea to save script code into automation rule, and in this case it will work more consistently.

Fibery Workspace Builder

It is a novel way to setup Fibery for novices.

Our naive approach to generate one single long script for a prompt did not work well. It lead to some critical problems

  • If execution fails, we re-construct the whole script again and
    • it lead to data duplication in selects, or data duplication in entities
    • it is slow, since script can be large
    • it is hard to roll back changes, since changes are huge
  • It is too easy to create not-valid schema (like add two titles fields, etc). Here we need stronger validation from Core part → Andrew can do it, but not now…

Here are 3 videos that demonstrate progress. You can watch just the last one if you wish to know the current state and don’t like history.

Then we discussed various ideas, and Aleh suggested to split huge script to blocks, execute blocks one by one and re-write only failed block. It immediately improved Build mode for Agent, removed duplicates in most cases and improved speed. From what I see, this is a very promising approach that we can push to production in near future.

Then we talked again and improved this flow, so the third version is even better. Faster, less errors, etc.

Next steps is to improve quality and speed of this builder mode and release it as an experimental thing. Then we will work in usability, undo, better UI and more use cases, like reports and automations.

You can try it for yourself already if you are brave.

Whiteboard

Only Vadim is working on Whiteboard from development point of view, so we have few features in nearest future plans:

We’ve released Whiteboard: Sections improvements and Upload multiple images We’re continuing fixing the behavior of basic elements and undo functionality, and we’re adding support for embedding views directly onto the whiteboard. After that, we’ll move on to working with templates.

Next

  • Whiteboard: make “Undo” work well
  • Embed views into the whiteboard (reports, tables, etc)
  • Whiteboard: Groups improvements
  • Whiteboard: Connectors improvements

Mobile

No progress made in Mobile

Files Fields

We’ve started requirements gathering and design phase for Files Fields re-work. Initial minimal scope is:

  • Allow to have many Files fields per database
  • Make great UI for Files on entity
  • Make great File units for all views
  • Add Gallery View

P.S. Ask your questions and provide insane ideas :alarm_clock:

8 Likes

Nice! Thank you for all this! We’re especially looking forward to the permissions, Gantt view, and files fields improvements! :heart:

2 Likes

Just a quick question

Will the file fields allow restrictions for file extensions? I know it’d be in name-only since you really can’t stop people renaming .bat or .exe to .png. Still nice to have if you’re expecting images to receive only images

No reason why this couldn’t cue up a proper content type peek using something like GitHub - openpreserve/libmagic-jna-wrapper: A Java/JNA wrapper for calling libmagic. :wink:

(post deleted by author)

This will be huge

But this might be huge-er.
Before deciding on Fibery, I tested a lot of tools, including some more niche ones. The only ones that felt like a paradigm shift and a significant gap in how people work (instead of some minor increment) were ones that had data-driven whiteboards.

The reason why I keep harping on about whiteboards so much is that from my experience collaborating with lots of people (devs, designers, clients, younger, older, more tech saavy or less tech saavy) the one commonality they have when it comes to whiteboards is that people just “get it”. It’s a flat canvas, a big piece of virtual paper which people know how to find their way around, and instead of relying on custom navigation and sidebars, people can find their way around the whiteboard by moving in a 2D space, and use it to memorize where things are (aha, the thing was above and to the right of this other thing).

As such, I’m hugely excited about embeding views in whiteboards as it could be a game changer.
My sugesstion to you @mdubakov and the rest of the Fibery team would be to take a look at https://www.creately.com/ and to a lesser extent https://weje.io/ as some potential tools to draw inspiration from when it comes to adding views to whiteboards.

One thing that creately had which made me go “holy shit, this is amazing” was that you could put multiple blank views / templates on a whiteboard (such as a timeline, or a effort/business value Matrix) and you could just drag entities over those views and they would adapt / transform to it.

For example, the same Epic entity would show up as a bar on the timeline view, but I could drag it over the effort/business matrix and it would show up as a circle in the correct position in the matrix.
I’m not saying Fibery should have that (though it would be amazing) but it might be something which sparks some interesting discussions in the team.

PS: Vadim for president, for developing the Whiteboard further :smiley:

3 Likes