Ember AI is an assistant app designed to help a user craft a well written, complete prompt to use when utilizing an AI platform in seeking information. It is based on the Pentagram method, or framework of crafting a prompt which is composed of five parts.
- The Persona defines a role that the LLM, or AI, should assume to give a point of reference for context of it's response.
- The Context gives background information for the LLM to use so that the desired task intent is understood.
- The Task should clearly define what the user is asking of the LLM and should include specific action verbs and instructions.
- The Output should describe the desired format of the response from the LLM.
- Constraints should list any boundaries or restrictions for the LLM response or describe desired tone if applicable.
To use Ember AI a user will first navigate from the landing page to the prompt input page.
On the prompt input page, the user will be guided through the five components of crafting a prompt. At each stage of the prompt creation, the user is required to make an entry before moving advancing. Information about what is intended at each phase is explained and an example given as a guide. Once the final phase, Constraints, is entered, the user will click the button to review the combined sections as an overall prompt.
Once on the Review page, a user will see the overall prompt created from the five sections they entered and has the ability to edit the prompt, if desired. Once the user is satisfied with the prompt, they can click the button to Generate Result which will submit the prompt to Gemini AI for a response, or they can choose to copy the prompt to paste in a different AI, window, application, etc.
If submitted on the page to the Gemini AI, the response will appear at the bottom of the page.
To run the app locally, from your command line interface (CLI):
git clone https://github.com/chingu-voyages/V54-tier2-team-23.gitcd V54-tier2-team-23npm i && npm run dev
Ember AI was built using:
- Framework: Next.Js
- Library: React
- Styling: Tailwind CSS
- Programming language: Typescript