Notes are dynamic, rich-text documents within your Workspace that serve as an interactive canvas for communication between team members and AI models. They provide a flexible surface for rendering information, from simple text to complex computational outputs.
Notes can be created and modified by both team members and AI systems within a Workspace:
Here's an example of a Worker that is set up to keep an up-to-date note on the current status of a project. It has its email address enabled, so that it can be copied into the project email threads.
[Placeholder: Example Worker]
As the Worker runs, it updates the Note with the latest information.
[Placeholder: Example Note]
You can similarly use Notes to store inputs and outputs to use in your Chat sessions. For example, you could maintain Notes that track the progress and details of individual prototypes, builds, or experiments. When starting a new Chat session, you can quickly bring the AI up to speed by asking it to read these Notes for context.
The AI can also help maintain these Notes by capturing key insights and decisions from your conversations. This creates an efficient workflow where Notes serve both as reference material for discussions and as a living record of your project's development.
Each note is composed of a series of nested blocks that can be created, modified, and updated both manually and programmatically. The following block types are supported:
Here's an example of a Note that demonstrates several of these elements, including a title, text, LaTeX equations, computation results, and code blocks:
[Placeholder: Example Note]
And here is what that Note looks like literally as a data structure for advanced users:
[
{
"type": "title",
"children": [{ "text": "Analysis Results" }]
},
{
"type": "paragraph",
"children": [{ "text": "Here are our key findings:" }]
},
{
"type": "latex",
"value": "f(x) = \\int_{a}^{b} x^2 dx",
"displayMode": "block",
"children": [{ "text": "" }]
},
{
"type": "compute",
"value": "8f47e8c1-a893-4d14-9925-c274def3859b",
"displayMode": "block",
"children": [{ "text": "" }]
},
{
"type": "paragraph",
"children": [{ "text": "As we can see from the computation above, the results confirm our hypothesis." }]
},
{
"type": "code",
"value": "import numpy as np\n\ndef analyze(data):\n return np.mean(data)",
"language": "python",
"children": [{ "text": "" }]
}
]
Coming Soon: