Pushing custom data to Speckle server

Features & Ideas
, ,
1

Hi, I am working mainly with ETABS and found out some miscellaneous data I need is missing (e.g story height/elevation, point coordinates, etc. Might be more in the future).

I want to ask is there a way for user to push their own custom data to Speckle server? so that I can fetch them to be used in another apps?

I found this specklepy SDK which states

create and extend your own custom Speckle Objects

So, my expectation is that I can build something that can be used to push my custom data (in specific format from ETABS or Excel) and holds it on Speckle server.

Can I use specklepy for that? If not, is there any alternative way?

1 Like
2

The standard Base object is dynamic, so you can shove custom data into any object like this:

obj = Base()
# add a primitive 
obj['some_param_name'] = 'spiffy custom data'
# or add an obj
obj['another_param_name'] = Base()

If you want to extend a speckle object, then something like this should work:

class MyCustomObj(Base, speckle_type="SuperAwesomeObj"):
    ...
3

Yes, use the published SDKs. That remains the correct way to publish and load data.

Key points

  • Versions are immutable; any change creates a new version.
  • Arbitrary data will store, query, and display fine online., in analytics tools like PowerBI or Speckle Intelligence
  • Host applications won’t act on it unless they already know how.
  • Custom fields must be JSON serializable.

C# and Python SDKs

  • C#: best for integrations or automation in .NET workflows.
  • Python (specklepy): best for scripting, analysis, or wrangling for onward integration.

Expectations

  • You can view and query anything you add.
  • Receiving applications will ignore unknown fields.
  • You should not expect to author native application objects.

Practical notes

  • Keep data lightweight and logically grouped.
  • Include identifying keys or IDs for joins.
  • Record units and coordinate context in plain fields.
  • Avoid tight coupling to authoring tool schemas; treat Speckle as a neutral data layer. Albeit with conventions.

Summary
Use C# or specklepy SDKs for stable publishing and retrieval. Versions are immutable, and host apps will only respect data they understand. Treat Speckle as your store of structured information, not a native object authoring system.

Beyond these generic points there are fine tuned answers that we’d need to know more concretely what you want to achieve before saying if it’s even possible.

Scope limits and guarantees

  • Will: your data is stored, queryable via GraphQL, viewable in web, retrievable via SDKs.
  • Might: downstream apps interpret your fields if they already understand them.
  • Will not: arbitrary fields force host apps to honor them; that needs explicit receiver logic.
  • Will not: guarantee persistence of arbitrary data in round tripping
1 Like
4

Hi @jonathon , thank you for your response, I have a look into C# SDKs, and found these speckle-sharp-connectors and speckle-sharp-sdk. It’s quite confusing for me, so can you suggest where should I start?

I mainly working with ETABS connector and I am fine with data coming from current connector. Sometimes I need specific information from ETABS model or even a table from Excel which is not available for default connectors (for now it is story elevation, joints/point coordinates, and I have high belief that I will need more in the future)

I am visioning that I can create something like a “mini connector” to add these custom data to my Speckle project and query them for my web application.

5

Ahh then thats perfect and well within scope of what is easily possible.

What is your experience with coding in general?

7

Hi @jonathon

Ahh then thats perfect and well within scope of what is easily possible.

Great to hear this before I go down this path

What is your experience with coding in general?

I can work with both Python and C#, but for the sake of .NET automation (which mainly the technology behind AEC application) I would like to go with C# (if this is what you are asking :hugs:)

8

The interesting thing about working with Speckle data is that you aren’t tied to the fact the original source would influence your tech stack. So you are free to use either SDK which will give you the tools to retrieve datas from the Speckle server and then its over to you.

Your Speckle server doesn’t currently support the infrastruture behind Speckle Automate or Intelligence else Id suggest either or both of these to work with.

I have looped in @bjoern into this as if there is ket information from ETABs we are not extracting but should then this can be potentially be Speckle action and then allow you to just work on your application.

9

Hi @jonathon,

I have various topics in the forum regarding data in ETABS with @bjoern and I have highlighted some data I need along the way.

While waiting updates from Speckle team, we want to create and try to push custom data first and see how it is integrated to our current app.

I am not very clear about Speckle Automate or Intelligence that you mentioned, but if that solve my current problem, I would be happy to use them, after all, I do not want to re-invent the wheels :innocent: