Inspiration
The creator economy is booming, but for Nano and Micro influencers, the reality is often a grind. We realized that while these creators have high engagement and authentic voices that brands crave, they lack the infrastructure of established agencies. We saw talented creators spending 80% of their time doom-scrolling for contacts and copy-pasting generic DMs, only to be met with silence. They want to create content, but they’re stuck acting as their own exhausted sales team. We asked ourselves: Why is the bridge between authentic creators and brands so manually intensive? We built Kollab to level the playing field. We wanted to give the underdog creator the same power as a full-service agency—an intelligent partner that knows who to talk to, what to say, and how to close the deal, so they can get back to what they do best: creating.
What it does
Kollab is an intelligent mobile AI agent designed to automate the business side of being a creator. It streamlines the entire partnership lifecycle, starting with Instagram.
- Smart Opportunity Scouting: Instead of guessing which brands might be interested, Kollab analyzes the influencer's niche and audience to identify high-probability brand partners who are actively looking for micro-influencers.
- Hyper-Personalized Outreach: The AI doesn't just send spam. It scans the brand's recent activity and requirements to generate tailored, high-converting outreach messages that actually sound like they came from a human who did their homework.
- Automated DM Workflow: Kollab handles the heavy lifting of the "cold call." It automates the DM process on Instagram, sending the initial pitch and managing follow-ups, ensuring no opportunity falls through the cracks due to a lack of time.
In short: Kollab is the dedicated sales manager that fits in a Nano influencer's pocket.
How we built it
The mobile app is built by flutter so we can provide end users seamless experience between mobile apps(both iOS and Android) and browser. We used instagapi and gmail api to connect users profile and help users to do their management.
In the backend, we use python fastapi to deliver all services. We utilize the google's gemini model to do all model related calls. In addition, we use Gemini-3-flash and browser use to build a Computer Use Agent to automate the Instagram data collection as well as bypass the strict bot detection in ins.
Challenges we ran into
The biggest challenge in this process is to bypass the Instagram bot policy where at the same time collect user information and get the ability to send direct message to any contacts.
In order to fix it, we implemented a MVP level CUA to help us connect with Instagram without worrying the anti-bot mechanism. This includes: login, searching, 2fa connection, DM, and view profile.
Accomplishments that we're proud of
The CUA agent that operates on Ins. The entire pipeline is fully automated.
What we learned
How to implement a customized browser use based CUA agent.
What's next for Kollab
We want to find potential uses and test our application. Keep building the end to end workflow and figure out the best way to connect multiple platforms for a user with the latest LLM. Use LLM/VLM to help users build profile.
Log in or sign up for Devpost to join the conversation.