lexica walkthrough + tutorial // creating ai-generated art with stable diffusion

kelly kim
16 Dec 202208:55

TLDRIn this walkthrough tutorial, the host introduces Lexicon, an AI tool likened to Google for AI art. The video demonstrates the search and generate functions, showcasing how prompts lead to AI-generated images. The host emphasizes the platform's potential for creativity and exploration, including the use of seeds for similar creations and the guidance scale for refining results. Tips for creating unique AI art are shared, highlighting the experimental nature of the technology.


  • 🎨 Lexicon is a stable diffusion search engine, often referred to as the 'Google for AI art'.
  • πŸ” Users can input prompts and receive AI-generated images based on those prompts.
  • 🌟 The platform showcases a variety of AI-created artworks on its homepage, similar to a social media feed.
  • πŸ‘ Users can interact with the artworks by liking them, just as on platforms like Pinterest.
  • πŸ’‘ The 'seed' of an image is a unique number that can be used to generate similar images.
  • πŸ“ The 'Dimensions' feature allows users to specify the size of the image they wish to generate.
  • 🎭 The 'Guidance Scale' adjusts how closely the generated image follows the input prompt.
  • πŸ” Users can perform reverse image searches, uploading an image to find similar AI-generated content.
  • 🎨 The 'Generate' function enables users to create their own AI art by inputting a description and adjusting settings.
  • 🌈 Combining different elements like medium, subject, and style can produce diverse and unique AI-generated artworks.
  • πŸ’Œ The presenter encourages users to share their AI-generated art and engage with the community on social media platforms.

Q & A

  • What is the AI tool being discussed in the transcript?

    -The AI tool being discussed is Lexicon, which is described as a stable diffusion search engine for AI art.

  • How does the AI tool Lexicon work?

    -Lexicon works by taking a prompt from the user and generating AI art based on that prompt. Users can search for existing AI-generated art or create their own using the platform.

  • What is the purpose of the 'seed' in Lexicon?

    -The 'seed' is a unique number generated when an image is created. It can be used to create similar images with minor alterations.

  • What is the 'guidance scale' used for in Lexicon?

    -The 'guidance scale' determines how closely the generated image adheres to the input prompt. A lower scale number allows for more diverse and potentially creative results.

  • How can users interact with AI-generated art on Lexicon?

    -Users can like, comment, and share AI-generated art pieces, similar to social media platforms like Pinterest.

  • What is the reverse search feature in Lexicon?

    -The reverse search feature allows users to upload an image and find similar AI-generated art based on that image.

  • What type of object does the speaker search for in the example given?

    -The speaker searches for sourdough bread and then specifically for an image related to the TV show 'Peaky Blinders'.

  • How does the speaker describe their approach to creating AI art?

    -The speaker uses a formula starting with the medium, followed by the subject, and then adding details and styles inspired by specific artists or art movements.

  • What is the speaker's suggestion for new users of Lexicon?

    -The speaker suggests starting with a lower guidance scale for more diverse results and encourages experimentation to find one's own tactics in creating AI art.

  • How does the speaker envision the use of AI-generated art pieces?

    -The speaker suggests using AI-generated art pieces for Christmas cards, e-cards, or even printing and displaying them at home.

  • What does the speaker invite users to do after creating their AI art?

    -The speaker invites users to share their creations on social media, tag them, or email them to see the diverse uses of the Lexicon tool.



πŸ–ŒοΈ Introduction to Lexicon: The AI Art Search Engine

The video begins with the host expressing gratitude for the audience's participation in another tutorial focused on an AI tool called Lexicon. Described as a Google-like search engine for AI art, Lexicon allows users to input prompts and receive corresponding AI-generated images. The host encourages viewers interested in AI to subscribe, like, and share the video. The tutorial starts with exploring the search function and then moves on to the generation function. The host clarifies they are not an AI expert and invites viewers to share tips in the comments. The homepage of Lexicon is showcased, displaying a grid of AI-created artworks. The host selects an artwork to delve into its details, such as the prompt, seed for generating similar images, and the guidance scale, which adjusts the accuracy of the prompt in the generated artwork. The video also demonstrates how to explore styles and iterations of different types of artwork based on a given prompt.


🍞 Exploring AI-Generated Art and Search Features

In this paragraph, the host continues the tutorial by illustrating how to search for specific subjects on Lexicon, using 'sourdough' as an example. The AI-generated results are showcased, and the host comments on the mundane nature of bread, prompting a search for a more specific subject, 'Pea Blinders,' a favorite TV show. The host finds an image of the character Tommy and discusses the show's attributes. A reverse image search feature is introduced, where the host uploads an image of Cillian Murphy to find similar AI-generated images. The host then transitions to the generation function, explaining the interface and how to input a prompt to create an AI artwork. The paragraph concludes with the host's intent to experiment with the tool and share their creations on social media platforms.

🎨 Creating AI Art: Techniques and Settings

The host delves into the process of creating AI art on Lexicon, starting with the input of a description, including keywords and negative prompts. Dimensions and guidance scale are discussed, with the host explaining the purpose of the guidance scale in achieving a desired level of accuracy in the generated image. The host shares experimental settings and introduces tactics learned from the build space community, emphasizing the fun aspect of creating AI art. A formula for creating AI art is presented, starting with the medium, followed by the subject, and additional details for a unique result. The host demonstrates this by creating an oil painting of a Boston Terrier in a psychedelic garden, inspired by artist Fernando potato. The paragraph concludes with the host encouraging viewers to create and share their AI-generated art, offering to see and appreciate their creations on social media.



πŸ’‘AI Tool

AI Tool refers to artificial intelligence software or applications that perform specific tasks using machine learning and cognitive functions. In the context of the video, the AI Tool being discussed is Lexicon, an AI platform for generating art based on user prompts. It is likened to Google but for AI-generated art, showcasing the advancement in technology and its creative applications.

πŸ’‘Stable Diffusion Search Engine

Stable Diffusion Search Engine is the core functionality of the Lexicon platform, which allows users to generate AI art by inputting text prompts. It represents a type of AI model that can create visual content based on textual descriptions, demonstrating the convergence of art and technology.


In the context of AI and Lexicon, prompts are textual inputs provided by users to guide the AI in generating specific types of art. These prompts act as instructions or cues for the AI, determining the theme, style, or subject matter of the generated artwork.

πŸ’‘Artwork Generation

Artwork generation refers to the process of creating visual art through AI, as demonstrated by the Lexicon platform. This involves the AI interpreting user prompts and producing original pieces of art, showcasing the potential of AI in creative fields.


In the context of AI art generation, a seed is a unique numerical value generated when creating an image. Seeds can be used to recreate similar images or as a starting point for new creations, allowing users to explore variations of a particular style or theme.

πŸ’‘Guidance Scale

The guidance scale is a feature within the Lexicon platform that adjusts the level of adherence to the user's prompt. A lower guidance scale value allows for more diverse and less predictable outcomes, encouraging experimentation and creative exploration.


In art, medium refers to the materials or techniques used to create an artwork. In the context of the video, selecting a medium like 'oil painting' for an AI-generated piece is part of the creative process, guiding the AI to produce artwork in a specific style.


The subject in art refers to the main focus or topic of an artwork. In the context of AI-generated art, specifying a subject provides the AI with a clear direction for the content of the generated piece.

πŸ’‘Artistic Style

Artistic style refers to the unique way in which an artist or a piece of art is characterized, often defined by specific techniques, colors, or themes. In the context of AI art generation, users can request the AI to emulate certain styles, such as 'psychedelic' or 'Arts nouveau,' to create distinctive pieces.

πŸ’‘Reverse Search

Reverse search is the process of searching for images based on a pre-existing image, often used to find similar images or the source of an image. In the context of the Lexicon platform, reverse search allows users to upload an image and receive AI-generated images that are similar in style or content.

πŸ’‘Creative Exploration

Creative exploration refers to the process of experimenting with different ideas, styles, or techniques to discover new approaches in art. In the context of the video, it emphasizes the importance of trying out various prompts and settings in the AI art generation process to achieve unique and inspiring results.


Introduction to Lexicon, a stable diffusion search engine for AI art, compared to Google for AI-generated images.

Exploration of the search function to find AI-generated images based on prompts, similar to social media platforms like Pinterest.

Understanding the concept of 'seed' in generating AI art, which is a unique number generated upon image creation and can be used for similar creations.

Explanation of the 'guidance scale' and its role in determining the accuracy of the generated image to the input prompt.

Demonstration of exploring different styles and iterations of AI-generated artwork based on a given prompt.

Searching for specific themes or objects, such as 'sourdough', and the ability to find AI-generated images related to these searches.

Reverse image search functionality to find AI-generated images similar to a provided image, demonstrated using an image from the show 'Peaky Blinders'.

Transition from searching to generating AI art, highlighting the creative process and interface.

Description of the 'negative prompts' feature, allowing users to exclude certain elements from the AI-generated images.

Discussion on the selection of image dimensions and the impact on the AI art generation process.

Sharing of personal tactics and strategies for creating AI art, influenced by the build space community and trial and error.

Presentation of a formula for creating AI art, starting with the medium, followed by the subject and additional details for customization.

Inclusion of specific artistic styles and details in the AI art generation process, such as 'cinematic lighting' and 'Arts nouveau'.

Demonstration of generating multiple AI art pieces and the potential for using them as Christmas cards or other creative applications.

Invitation for the audience to create their own AI art and share it on social media platforms, fostering a community around AI-generated art.