Local Providers

Local providers in Scribeist

Scribeist supports these local BYOK providers:

• Ollama
• LM Studio

These providers are useful if you want to run models on your own machine instead of paying a hosted API provider.

How local provider models work

• Ollama shows the models available from your Ollama instance
• LM Studio shows the models available from its local API server

This means Scribeist only shows models your local setup is actually serving.

Connect Ollama

1. Ollama

Make sure Ollama is installed and running on your machine.

The default local endpoint is usually:

http://127.0.0.1:11434

3. Pull a model

Example:

ollama pull qwen2.5:0.5b

4. Connect it in Scribeist

1. Go to Account > Integrations
2. Open the Ollama connection
3. Enter the base URL, usually http://127.0.0.1:11434
4. Save

After the connection succeeds, Scribeist lists the Ollama models available from that server.

Connect LM Studio

1. Download a model

Downloading a model is not enough by itself.

The model must also be loaded into LM Studio and exposed through the local server.

2. Open the Developer tab and load the model

Open LM Studio and select the Developer tab. Choose the model you want Scribeist to use and load it into memory.

Downloading a model is not enough—it must be loaded in the Developer tab before LM Studio can serve it.

3. Start the local server

In the Developer tab, turn on the local server using the server button.

4. Enable CORS

Select the Server Settings button beside the server button, then turn on Enable CORS. This allows the Scribeist website to connect to LM Studio from your browser.

The default local endpoint is usually:

http://127.0.0.1:1234

5. Connect it in Scribeist

1. Go to Account > Integrations
2. Open the LM Studio connection
3. Enter the base URL, usually http://127.0.0.1:1234
4. Save

After the connection succeeds, Scribeist lists the models currently exposed by the LM Studio server.

Quick test

If you want to confirm the local server is working before connecting Scribeist:

Ollama

Open:

http://127.0.0.1:11434/api/tags

LM Studio

Open:

http://127.0.0.1:1234/v1/models

If you get a valid response listing models, the server is running.

Troubleshooting local providers

I downloaded a model but Scribeist shows nothing

For local providers, downloading is not always enough.

Check:

• the local server is running
• the base URL is correct
• the model is actually loaded or available from that server

The connection fails to save

Check:

• the local app is running
• the endpoint URL is correct
• no firewall or local network rule is blocking the request
• the local server allows requests from https://app.scribeist.com

I see a CORS error or 403 Forbidden

The local server is running but has rejected the Scribeist website's browser request.

For Ollama, quit every Ollama process and restart it with OLLAMA_ORIGINS=https://app.scribeist.com. Merely closing the Ollama window may leave its background process running.

For LM Studio, check its local server settings and enable cross-origin access for the Scribeist website.

The port is already in use

Only one server can listen on a local port at a time. Fully stop the process already using port 11434 for Ollama or 1234 for LM Studio, then start the server again.

I only use local providers. Will Scribeist still work?

Yes. On a BYOK plan, Scribeist can use Ollama or LM Studio without needing a hosted provider connection, as long as the local provider is reachable.