Qalam AI

Inspiration

I believe that at least once in our lives, we’ve said, “I WANT TO WRITE A BOOK!” We start with great motivation, but soon after, we get stuck. Qalam AI is designed to help those who find themselves at a dead end in their writing by asking leading questions. Our goal is to guide writers back into the flow of writing without taking away their creativity, enabling them to overcome writer's block, whether they’re writing a novel, screenplay, or blog post.

What it does

Qalam AI acts as an intelligent writing assistant that engages with writers who feel stuck. The web app prompts users with context-aware, leading questions based on the story or content they’ve written so far. It detects key elements like characters, locations, or events and generates helpful suggestions or questions to guide writers. Rather than providing answers or writing for the user, Qalam AI enhances the creative process by stimulating fresh ideas and perspectives.

How we built it

We built Qalam AI using Reflex for app development, which allowed us to quickly create the user interface. We used three large language models (LLMs) deployed via Groq: one to identify objects like characters and locations, one to summarize input and update a database, and the third to generate suggestions based on the user’s story. For voice generation, we integrated Cortesia, and we used SingleStore as our database solution to track and store story elements.

Challenges we ran into

One of the biggest challenges we faced was ensuring that the AI didn’t overstep by taking away creativity from the writers. Balancing useful suggestions with maintaining the writer’s control over the story required fine-tuning our models. Another technical challenge was learning how to use Reflex and Groq on the fly, as neither were familiar to us prior to starting this project.

Accomplishments that we're proud of

We’re proud of successfully creating a tool that can help writers without stifling their creativity. One of our proudest moments was seeing Qalam AI ask thoughtful, contextually accurate questions that made users pause and think, rather than giving them solutions outright. It also felt great to build something from scratch, learning new technologies and frameworks along the way.

What we learned

Through this project, we learned how to work with Reflex and Groq, as well as the intricacies of implementing large language models in a way that supports creativity rather than hinders it. We also learned a lot about user experience and the importance of creating a product that feels like an extension of a writer’s thought process, rather than a disruptive tool.

What's next for Qalam AI

In the future, we plan to integrate more sophisticated personalization features, allowing users to adjust the tone and depth of the suggestions they receive. We also want to enhance the voice generation capabilities so that Qalam AI can offer auditory prompts to those who prefer to listen to ideas. Additionally, expanding to more languages and genres is a key goal, ensuring that writers of all types can benefit from the tool.

Built With

• cartesia
• grog
• python
• reflex
• singlestore