Cannot Connect to Custom API on Localhost

Plasmic Studio Questions
1

I’d like to be able to use my own self-hosted API to make MS SQL calls to our on-premise database. I tried locally installing plasmic, but it can’t access the local IP address for some reason as a data source. I curled from the same machine, and it got back the json response, so it isn’t an env thing. My API only ever received responses from the command prompt, not the local plasmic.

That’s strange enough, but when I try the REST API Component on the Plasmic Studio cloud app… it also never makes a call to my locally hosted API, and I thought it could?? It says it is browser-based calls. My browser can go to the GET URL and out pops the data displayed on the screen.

I know there are custom code components, but the way it sounds… these too will have problems, since it seems to require public app hosts. I can’t exactly share our data with the outside world, nor does the company provide me the means to expose endpoints to the world. Red tape city.

The API I created is just a simple node app that generates API endpoints for MS SQL data+view tables. Our internal AI made it for me in like a day. I made the repo public for now if you’re wondering if this is the problem:
EmCmEdT/MsABON: MsABON automatically discovers MS SQL tables and views matching configured filters and exposes REST endpoints for them.

Though it seems fine to me.

The original app was designed and deployed in Budibase, but IT is not supporting a local installation for containerized apps, and my department only supports local installation, so I have to migrate to something that can be deployed locally on Windows, like Plasmic apps.

I am in a pickle, because I really would like a no-code solution to ease in the non-dev team to design or debug on my behalf, but I’m so close to going full code and scrapping my dream.

Frustrated.

2

Hi @EmCmNeDt, you can use your new API, but Plasmic servers would not have access to the API on your local machine. You will need to host it somewhere and make sure it’s accessible to the internet.

3

Is a fully on-premise solution not possible?

It’s a security issue to make our data accessible from the web, even properly authenticated. Suppose I were to fully migrate to a cloud DB, which is the cleaner option vs. a script that transfers data between the two. I’m not sure who I need to convince that this is okay, and we’re talking about new system validation and a deviation from standard, so the potential for a second kibosh is high. On-premise SQL Server is already setup and is fully validated.

4

Okay, I figured it out.

I can indeed setup a localhost app host. When I did that, the studio.plasmic.app’s HTTP REST API Fetcher component started magically working.

Here’s the short version of how I got it to work:

  1. Published my app from the studio.plasmic.app
  2. With node installed (used the one in .tool-versions), ran npx create-plasmic-app@latest in a folder of my choice.
  3. Configured the app. Chose default everything.
  4. Ran npm run dev
  5. In plasmic studio, hit the triple dots in the project dropdown, selected configure custom app host.
  6. Changed https to http, and typed out localhost:3000/plasmic-host
  7. Success

Source 1: Cannot make integration API calls - Plasmic Studio Questions - Plasmic Community

Source 2: New Plasmic codebase (CLI) | Learn Plasmic

Source 3: Host Plasmic Studio in your app | Learn Plasmic

Originally, my app wasn’t published, but once I cleared the app host, published, and selected http instead of https, it worked.

1 Like