Influsend

Inspiration

Facilitate influencers and public figures to access job offers made to them through social networks, specifically through Instagram

What it does

Classify messages received on Instagram, save those related to work and are displayed on the web.

How we built it

First, we connect to the user by accessing the Instagram API. At this time we collect the messages you review. Secondly, we classify the recurring messages, selecting those related to work. We connect with the OpenAI API to bias the messages, interpreting them and labeling them with gpt-4-o-mini. Finally, messages related to work are displayed on the web, along with the time they were sent and the company that sent them.

Challenges we ran into

reply: AI Agents

What we learned

We have learned to work with APIs, digital agents, and databases, which has been an incredibly enriching experience. We are especially thrilled to have had the opportunity to collaborate and experiment with OpenAI technology, which has expanded our capabilities and understanding in developing advanced solutions. This process has been inspiring and motivates us to continue exploring the potential of these tools!

What's next for Influsend

Influsend aun tiene mucho recorrido, a continuacion explicaremos una serie de possibles mejoras futuras:

AI-Generated Responses:On the website we have created a button to generate an automatic response to the offer. The button function is not yet generated. The idea is to create a button that allows the user to generate an automatic response. This feature is intended to save time by automatically creating personalized replies based on the message content and context.

Separate job offers by priority: Differentiate offers based on the money they provide to the user. In this way we can sort them and show the most interesting and useful ones first.

Broader classification: The next step would be to train the model to classify more than one type of message. The idea would be to allow the user to determine the labels, for example: family, friends, work, etc., and generate a tab with the messages of each category separately. In this way we would have all the messages in to one classe, separated according to the user's priorities.

Built With

• astro
• ccs
• html
• instagram
• javascript
• openai
• openaiapi
• python
• sqlite