MemoAgent — Simple overview

Demo: https://youtu.be/e4BuAgaFtPM

---

Project Overview

MemoAgent is an intelligent, conversational video assistant. Upload any video and ask for edits, highlights, summaries, clips, or audio fixes in plain language — no editing skills required. It saves user preferences so future jobs follow a consistent style. MemoAgent helps creators, students, marketers, and teams save time and produce polished short videos from long footage.

---

Team Introduction

I am a student from Bangladesh.
Role: student

---

Key Features & Tech Stack

Key features

• Natural-language requests for video edits (e.g., "Make a 30s highlight reel").
• Summaries & scene search (find where topics/words appear).
• Clip extraction for social formats.
• Audio clean-up: noise removal and voice boost.
• Memory-aware personalization (remember preferred styles).

Tech stack

• memories.ai API — stores and recalls user preferences, saved styles, and past jobs.
• runwere.ai API — runs video processing tasks (trimming, filters, assembly).
• Web UI (uploader + request box) — user-facing interface for uploads and plain-language requests.

---

Sponsor Tools Used (and how they were integrated)

• runwere.ai — Integrated as the primary video processing engine. After MemoAgent parses a user request into processing steps, it calls runwere.ai APIs to extract clips, apply filters/transitions, and assemble the final video.
• memories.ai — Used for persistent memory: saving user style presets and retrieving them to personalize subsequent jobs.

---

Challenges & Learnings

Challenges

• Converting open-ended natural-language requests into reliable processing steps.
• Handling very large video files while keeping processing time reasonable.
• Ensuring privacy and safe handling of uploaded content.

Learnings

• Specific, well-structured prompts (time ranges, mood, style) greatly improve results.
• Storing style presets yields more consistent outputs across multiple videos.
• Leveraging specialized APIs (runwere, memories) speeds development and offloads heavy processing.

---

Future Improvements / Next Steps

• Add preset styles (social, cinematic, podcast) users can select.
• Improve upload and processing speed with progress indicators for large files.
• Add stronger privacy controls and one-click deletion of uploaded content.
• Multi-language support for prompts and transcripts.
• In-app tutorial and a copy/paste prompt cheat-sheet for non-technical users.

---

How to use

1. Open MemoAgent (web app).
2. Upload your video file.
3. Type a request in plain language (e.g., "Create a 15s social cut with captions and upbeat music").
4. Submit and wait — processing time varies.

---