Comment on page
Collimator Chat, powered by ChatGPT, is currently in beta. We are constantly working to improve and add features, but this feature may still give inaccurate or misleading information, or simply fail to work under certain circumstances.
You can access the chat feature in the Model Editor, by clicking on the Chat button in the bottom bar.
The Chat panel is accessible through the bottom toolbar in the Model Editor
The Chat panel will slide upward, revealing a now-familiar chat interface.
You can type a question into the prompt entry field, or you can click one of the quick prompt buttons. The AI agent will process your prompt and generate a helpful response.
The Chat feature is free for all Collimator users to try. Everyone gets 10 prompts to see if the feature is useful for them. After that, you will have to create an OpenAI account, if you don't already have one, and navigate to https://platform.openai.com/account/api-keys to generate a secret API key. The Chat agent will prompt you to add your secret key to your account in Collimator. Keep in mind that you will need to paste the key you got from OpenAI into a secure text field in Collimator. You will only have one opportunity to copy the key from OpenAI's site. After that, if something goes wrong, you'll have to generate a new one. Secret API keys are free to generate, but using the Collimator Chat feature may incur charges with OpenAI, depending on your account type and usage level.
You can always manage your OpenAI keys (or any other 3rd party credentials) within the Collimator app from your Dashboard. Just look for the key icon in the top right corner:
We are working on adding to this list all the time, but currently you can use the Chat feature to:
- Ask questions about a (sub)model, including how it works or what it does.
- You can ask for the equations that underlie a model or submodel.
- You can run a simulation and plot the results.
- Each model or submodel has its own Chat context per user. When you leave Collimator and come back, you can pick up where you left off in your chat.
- You can also restart a conversation at any point, resetting the Chat agent's context to the default.
- If the response to your prompt is taking too long, you can 'Stop Generating'
- If something goes wrong, or you think a response you get could be improved, you can 'Repeat the Question'
- If you'd like the agent to expand on the last response, you can 'Keep going'