ScribeGenie

Empowers organizational continuity by crafting tailored grant/sponsorship proposals. It adapts to your format, tone, and past insights, ensuring seamless alignment with your identity and standards.

Comment

Inspiration

We work as a volunteers at Bright Minds Enrichment, a non-profit dedicated to accessible and innovative educational programs.It relies on grants, funding, and external sponsors to drive the mission forward. Crafting compelling proposals is integral to its operations. With a team of approximately 30 individuals, there is a lot of importance on adhering to each grant's unique requirements. Despite having templates in place, the diversity in writing styles among team members often extends the submission process to about a week at the earliest. This prolonged timeline is largely attributed to the manual iterations required to ensure alignment with our organization's established format, tone, and language.

What it does

Here's where Gen AI steps in to revolutionize our workflow. By harnessing the power of artificial intelligence, Gen AI streamlines our proposal development process, significantly reducing the time required to less than an hour. This accelerated turnaround not only enhances the efficiency but also allows organization focus more resources on the core mission of enriching educational opportunities for all.

ScribeGenie' crafts meticulously tailored grant or sponsorship proposal documents. Leveraging Gemini API, it intuitively adapts to the format, tone, and language established by your organization. Drawing insights from your past proposals and organizational profile, 'ScribeGenie' ensures seamless alignment with your unique identity and standards, facilitating efficient and consistent communication of your vision and objectives.

How we built it

We utilized Python with the Streamlit framework. Our system scrapes URLs for data and parses PDFs for organizational details and past correspondences. Engineering integrates Gemini Pro Google AI API, and LangChain for accurate data extraction. Ultimately, we generate tailored proposals seamlessly

Challenges we ran into

  1. Ensuring the generated proposal does not fabricate information or hallucinate: So had to sharpen the prompt engineering bit.
  2. Reading the PDFs and picking contextual information
  3. Ability to export the generated text as a PDF. It had to be converted and the text needed to be wrapped so that we could download it as a PDF file.

Accomplishments that we're proud of

Ability to take up this challenge and build something that would be useful to non-profits like Bright Mind Enrichment, other research or small businesses.

What we learned

Learnt to work with langchain, streamlit, generative ai applications, and FPDF

What's next for ScribeGenie

Creating a workspace like steps, so that the admin can be in control of adding information about the organization and past proposal letters. Then the organization members can be added by the admin and they can create proposals just with 1 click which is adding the URL with details about the grant or sponsor fund for which they want to create a proposal.

Built With

Share this project:

Updates