> ## Documentation Index
> Fetch the complete documentation index at: https://docs.typingmind.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Ollama (Local)
[Ollama](https://ollama.com/) is an open-source project that allows you to run many LLM models locally on your device.
You can easily config Typing Mind to run with Ollama, below is the full guide:
This instruction is for the TypingMind Web version ([https://www.typingmind.com](https://www.typingmind.com)). For the macOS version and Setapp version, due to Apple's security policy, requests to `http` protocol are blocked. If you want to connect to the macOS app, you can still follow the instructions here, but with one additional step: you need to set up HTTPS for Ollama. This can be done using various techniques (e.g., using a local HTTPS proxy). For more details on how to run Ollama on HTTPS, please reach out to the [Ollama project](https://github.com/ollama/ollama) for support.
## Download Ollama
Go to [https://ollama.com/](https://ollama.com/) and download Ollama to your device.
## Set up Ollama environment variables for CORS
Run the following commands so that Ollama allows connection from Typing Mind.
```bash theme={null}
launchctl setenv OLLAMA_HOST "0.0.0.0:11434"
launchctl setenv OLLAMA_ORIGINS "*"
```
**Important:** After running the commands, **completely kill the Ollama app and reopen/restart it** for the changes to take effect — **not just close and reopen it**.
Once the set up is done, run the following command to start a model (in this example: `phimini-4` )
```bash theme={null}
ollama run phi4-mini
# Output:
>>>
>>> Send a message (/? for help)
```
## Add Ollama models to Typing Mind
Open Typing Mind and open the Model Setting button, then click “Add Custom Model” on the top right corner. Enter the details as following: (in this example, we use phi4-mini)
* Model Name: Phi 4 Mini
* Model ID: phi4-mini
* Endpoint: [http://localhost:11434/v1/chat/completions](http://localhost:11434/v1/chat/completions)
* Click Test & Save so the app can automatically check the model's capabilities.
* Click Confirm & Proceed to add model.
## Chat with Ollama
Once the model is tested and added successfully, you can enable the model in model list. Then select the custom model and chat with it normally.
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.
## Set up Ollama environment variables for CORS
Run the following commands so that Ollama allows connection from Typing Mind.
```bash theme={null}
launchctl setenv OLLAMA_HOST "0.0.0.0:11434"
launchctl setenv OLLAMA_ORIGINS "*"
```
**Important:** After running the commands, **completely kill the Ollama app and reopen/restart it** for the changes to take effect — **not just close and reopen it**.
Once the set up is done, run the following command to start a model (in this example: `phimini-4` )
```bash theme={null}
ollama run phi4-mini
# Output:
>>>
>>> Send a message (/? for help)
```
## Add Ollama models to Typing Mind
Open Typing Mind and open the Model Setting button, then click “Add Custom Model” on the top right corner. Enter the details as following: (in this example, we use phi4-mini)
* Model Name: Phi 4 Mini
* Model ID: phi4-mini
* Endpoint: [http://localhost:11434/v1/chat/completions](http://localhost:11434/v1/chat/completions)
* Click Test & Save so the app can automatically check the model's capabilities.
* Click Confirm & Proceed to add model.
## Chat with Ollama
Once the model is tested and added successfully, you can enable the model in model list. Then select the custom model and chat with it normally.
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.