komo

WHERE IS THAT FILE??? Was it a thread in Slack? Or an import in Notion? komo is here to organize your chat histories and threads for easy retrieval. With komo, easily get into a "flow state".

Comment 1

komo helps you get into it!

Inspiration

As a busy working professional at a growing startup in San Francisco, Alvin has to deal with a ton of data every day to cross reference. Whether it's reallyyyyy longgg slack threads or files imbedded in notion, searching for data takes up so much time. Lots of data gets lost in company work softwares which decreases productivity.

What it does

komo is designed to optimize workflows by seamlessly integrating different workplace platforms together. It's a unified, intelligent workspace that declutters your digital life by letting you search for related documents, messages, and media across multiple platforms—all in one place. Instead of jumping between Slack threads, Notion pages, files, screenshots, and random PNGs buried in conversations, our tool becomes your One-Stop Search Hub. With a single query, it fetches content from your connected services, analyzes it, and presents everything in a clean, organized interface. Along with a search, it provides detailed summaries of where your query was referenced, in what context, and on which service/app. 1. User searches query: “Onboarding slides” 2. Backend performs: • Slack Search API • Notion Search API 3. Each service returns JSON metadata 4. If the result is a file (PPTX, PDF, image): • We call the platform’s file download endpoint (maxing API calls!) • Gather information (texts/messages) around the file to gain contextual references about the file 5. OCR output (clean text) is sent to an AI model for summarization 6. Frontend displays: • Unified Search Results • File/text preview • AI summary

How we built it

Frontend • React + TypeScript — Fast UI development with predictable component structure • Next.js — Super-fast dev environment w/ deployment features • Tailwind CSS — Rapid styling and responsive layout Backend • Node.js + Express — Lightweight, hackathon-friendly API server and routes • Custom /search endpoint — Unified search handler for all services • Internal service wrappers (not exposed): • notionService.ts — Notion Search + Page/Block parsing • slackService.ts — Slack Search + File retrieval • Gemini AI integration to summarize search!

External APIs & Integrations • Universal search • Slack Web API • Search messages/files • Notion API • Retrieve page content + file blocks

Challenges we ran into

Connecting API backend to our frontend was very tricky because we were working on different branches that weren't constantly updating. Also, fetching backend calls from frontend took a while to navigate. SlackAPI: documentation wasn't thorough, had to learn how to work with refresh tokens, API results didn't give us the info we wanted so we had to parse data differently. NotionAPI: searches took too long to query, we had to think of loop holes: stored cached db locally so that we didn't have to re-search every single page.

Accomplishments that we're proud of

We had a beautiful design from our PM who had never touched design before. She did amazing and the UI looks wonderful. We also have never used Slack and Notion API's, with limited documentation, we were able to seamlessly integrate it into our app through the countless Auth Tokens and Client IDs. Optimizing search times was also one of the biggest feats — our backend dev was able to maximize efficiency while minimizing API load. The goal of bringing down our search times by 96% also helped us not get rate limited!

What we learned

New technologies are always fun to try out and even with limited docs, no issue is irresolvable. With adequate research and concentration, each of our bugs were resolved. We learned how to access different workspaces through a common web app and how to leverage AI to help solve bugs but not take over the whole project.

What's next for komo

  1. Full Multi-Platform Expansion Right now we support Slack and Notion, but komo is designed to scale. Next integrations include: • Google Drive & OneDrive • iMessage/WhatsApp/Discord • GitHub issues & PR search • Email (Gmail, Outlook) • Open Source Extension based Developer kit
  2. Real Deployment: • Leveraging Cloud Services to reach real users • Storing compute instances and DB setup to manage data • Optimizing search and UX based on feedback
  3. Real Authentication (OAuth) During the hackathon we used hardcoded tokens for speed. Next steps: • Implement secure OAuth for Slack + Notion • Add token refresh + encrypted storage • Let users manage connected apps in a “Linked Accounts” panel
  4. Smarter AI Search + Summaries We want to develop deeper intelligence: • Automatically detect topics from your workspace • Multi-document summarization (“summarize everything about X”) • Timeline view of past conversations + files • Entity extraction (names, dates, diagrams)
  5. Visual & Media Intelligence We plan to extend support for: • Slide decks (PPTX → text + slide previews) • Videos (frame-based OCR + transcript search) • Tables and charts (AI extraction & interpretation)
  6. Custom Workspaces + Hotkeys We want to enable: • User-defined “workspaces” for topics like school, work, hackathon • Global hotkey (e.g., Cmd + Shift + K) to open komo anywhere • Spotlight-style popup search on your desktop

komo aims to be a one-stop-shop to search through your digital life! From searching through all your services, to boosting productivity by unifying your contextual-search, komo aims to be around by just being a shortcut away ;)

Built With

Share this project:

Updates