Skip to main content
The DinoAI prompt input is your main interface for interacting with the DinoAI copilot agent. Beyond typing your prompts, the input toolbar gives you quick access to a set of controls that shape how DinoAI responds and behaves.

​
Toolbar Controls

​
Attach Context

Click the @ icon to attach context to your prompt, such as files or folders from your project. This helps DinoAI understand the specific parts of your codebase you are referring to, leading to more accurate and relevant responses.

​
DinoAI Settings

Click the Settings icon to open the DinoAI Settings panel, where you can manage tool permissions and connected integrations for your DinoAI agent.

​
Chat History

Click the Clock icon to open your DinoAI Chat History, where you can browse and reopen previous conversations with the agent.

​
Credits and Context

The ring next to the approval mode picker shows how full the chat’s context is. Click it to see:
  • This chat: the credits this chat has used.
  • Used this month: your DinoAI credits this month.
  • Context window: how much of the chat’s context limit is in use, in tokens. When it fills up, DinoAI compacts the chat automatically.
The DinoAI prompt input with the credits and context card open: the credits this chat used, the credits used this month, the context window at 9% of 200k tokens, and the Compact this chat button
From 80% full, the ring turns red and shows the percentage. To compact a long chat sooner, select Compact this chat: DinoAI folds the older messages into a recap and keeps answering with that context. Sending pauses until it finishes, and the recap uses credits. The button is available once a chat is large enough to be worth compacting and DinoAI isn’t responding.

​
Approval Mode

Click the approval mode picker, which shows your current mode, to choose how often DinoAI asks for your approval before it acts in Agent mode: Manual, Auto, or Full auto. See Approval modes for what each mode runs without asking.
In Full auto, DinoAI runs every action without asking, terminal commands included, and you are responsible for what it runs.

​
Mode Selection

The ∞ Agent dropdown lets you switch between two modes of interaction:

​
Model Selection

The model dropdown (displayed as GPT-6.1 Sol by default) allows you to choose which underlying AI model powers your DinoAI session. Different models may vary in speed, capability, and suitability for different types of tasks. See Supported models.
The DinoAI model picker in the Code IDE, grouped by provider: Anthropic, Google, xAI, Moonshot AI, Qwen, DeepSeek, and OpenAI, with GPT-6.1 Sol selected

​
Suggested Actions

Below the prompt bar, DinoAI surfaces a set of Suggested Actions — one-click prompts that are dynamically generated based on the files you currently have open in the Editor.
Rather than typing a prompt from scratch, you can click any suggested action to instantly send it to DinoAI as a prompt. Suggestions update as you switch between files, ensuring they stay relevant to whatever you are working on at any given moment. Examples of suggested actions include:
  • Explain this model — get a plain language explanation of the dbt™ model currently open
  • Document my dbt™ model — automatically generate documentation for your model
  • Refactor dbt™ model — get recommendations or apply refactoring to your dbt™ model
  • Debug failing dbt™ model — diagnose and fix issues with a failing dbt™ model
  • Write unit tests — generate unit tests for the model in focus
  • Create .dinoprompts — create a .dinoprompts file to save and reuse custom prompts in your project
Suggested actions are context-aware. For the most relevant suggestions, make sure the file you want to work on is active and open in the Editor.
  • .dinorules — Set rules that apply to all Copilot interactions
  • .dinoprompts — Save and reuse prompts across your team
  • DinoAI Settings — Configure Copilot behaviour and model preferences