Features

Manage Chat Models

Control the use of chat models

Control the use of chat models

There are more than 10 chat models available on TypingMind custom:

  • GPT-3.5-based: GPT-3.5, GPT-3.5-16K, GPT-3.5 (0301)
  • GPT-4-based: GPT-4-8K, GPT-4-32K, GPT-4 (0314), GPT-4-32K (0314)
  • Claude: Claude 1, Claude 2, Claude Instant, Claude Instant 1.2

Personalize your chat instance to match your brand identity

Personalize your chat instance to match your brand identity

If you want to use your own LLMs, Azure OpenAI, OpenRouter or your private LLM, you can also add Custom Model (as long as it serves via an OpenAI-compatible API endpoint)

Personalize your chat instance to match your brand identity

With these models, you have the option to:

  • Pre-configure the chat model API key
  • Choose which models should be used for your users on the interface
  • Enforce users to use 1 model only with training data

Pre-configure the chat model API key

TypingMind allows you to pre-configure the chat model API key, such as OpenAI or Anthropic

Simply go to API Keys menu (under the General section) and enter the API key.

Pre-configure the chat model API key

From the chat UI, your members can instantly use the chat models without manually setting it up again:

your members can instantly use the chat models

Choose which models should be visible to users on the interface

You have full control over which specific chat models are visible and accessible to your users:

  • Navigate the “Manage Models” menu
  • Turn on / off specific models

Choose which models should be visible to users on the interface

This will directly affect the interface, the chat models which are turned off will be invisible and can’t be selected by your users

the chat models which are turned off will be invisible

In case you want to control the model availability for a certain group of users, please refer the Model Availability section at Usage & Limits.

Enforce users to use 1 model only with training data

If you believe that a specific chat model, such as GPT-4, is most compatible with your training data and can generate the best results, you can configure settings to make that model the default choice for users.

By doing so, there will be no option to change the model from the chat interface. This can be done by:

  • Navigate the Training data menu
  • Switch the toggle 'Enforce Default Model Settings' under the Model Parameters section to enable this option

Enforce Default Model Settings

Users won't have the option to select chat models from the chat interface:

select chat models from the chat interface

These options provide you with a high degree of control over the chat models available to your users, enabling you to tailor the chat experience to your specific needs and preferences.