JessyTsui

JessyTsui / cerul

Public

AI-agent video search — search, retrieve, find inside videos, not just tags and transcripts.

cerul.ai ai-agent api computer-vision multimodal neon-postgres
37
3
89% credibility
Found Mar 10, 2026 at 30 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Cerul is an early-stage open-source project building video search infrastructure that lets AI agents find visual content like slides and charts in videos, beyond just transcripts.

How It Works

1
🌐 Discover Cerul

You stumble upon Cerul while looking for ways to help AI assistants understand videos better than just listening to words.

2
🏠 Visit the homepage

Land on the welcoming site that explains searching slides, charts, and demos inside videos.

3
🎥 Try the live demo

Play with the interactive search box to instantly see video clips matching your questions.

4
📚 Explore guides

Read simple stories and examples about finding quick clips or deep insights from talks.

5
Pick your search style
📹
Fun clips

Grab ready-to-use video snippets for your projects.

🧠
Deep talks

Uncover key moments with visuals from lectures and shows.

6
👤 Set up your space

Sign up quickly to get your personal dashboard for tracking searches.

✅ Videos come alive

Your AI now pulls exact clips, timestamps, and visuals to answer questions perfectly.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 30 to 37 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is cerul?

Cerul is an open-source AI agent video search tool on GitHub that indexes videos for visual content like slides, charts, and demos, going beyond transcripts to make on-screen elements queryable via API. Developers get a unified search endpoint for two tracks—lightweight b-roll from stock sources like Pexels and knowledge retrieval from talks—with curl-ready requests returning direct video URLs, timestamps, and optional AI summaries. Built in TypeScript with Next.js frontend and Python FastAPI backend, it deploys easily to Vercel for agent integrations.

Why is it gaining traction?

It stands out among ai agent github repos by solving video analysis gaps for agents, delivering semantic matches like "AGI timeline slide" with evidence citations that transcript search misses. The live demo console on the landing page lets you test ai agent video examples instantly, and the shared API across tracks keeps integration simple without custom pipelines. Open-core approach shares the scaffold while hiding production indexes, hooking devs building ai agent video generation or research tools.

Who should use this?

AI engineers crafting agent workflows needing video RAG, like querying keynotes for visual claims in research bots. Demo makers sourcing free ai agent video b-roll for prototypes without manual curation. Contributors to ai agent github projects experimenting with open-source video search beyond text.

Verdict

With 16 stars and a 0.9% credibility score, Cerul is bootstrap-stage but boasts pro docs, mock APIs, and Vercel deployment—solid for early ai agent video repos. Watch or fork if video beats text search; skip for production until indexes land.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.