- Home
- AI Video Search Tools
- MomentSearch
MomentSearch
What it is
MomentSearch is visual video search and retrieval-augmented generation software that finds exact moments matching what appears or is said in a video. You can upload videos or paste YouTube URLs for indexing. Each search checks both visual frames and transcript text, then combines the results. Answers can point to clickable moments with keyframes, timestamps, synchronized transcripts, seen-or-said labels, and match scores. If you do not provide an LLM key, MomentSearch still returns ranked clickable moments but does not generate prose answers. A confidence gate can also decline to call an LLM when neither search branch meets its threshold. Models for answers, embeddings, reranking, and speech-to-text can be selected by provider name, with hosted and local options documented. The project includes Docker entrypoints and deployment guidance for Fly, AWS, and Google Cloud. It is free and open source under Apache 2.0, with API, Linux, macOS, Windows, web, and self-hosted options. Note that its API has no authentication and is single-user by default.
For and against
For
- Search combines visual frames and transcript text.
- Citations include keyframes, timestamps, and match scores.
- Ranked moment retrieval works without an LLM key.
- Models can use hosted providers or local options.
Against
- The API has no authentication.
- The application is single-user by default.
- SocialKit YouTube ingestion becomes paid after 20 free credits.
- Cookie authentication expires approximately every two to three weeks.
The verdict
MomentSearch combines visual and transcript search to surface specific video moments with clickable citations. Its lack of API authentication and single-user design are important constraints for deployment.
Who it suits
MomentSearch suits people who need to find and cite specific moments in indexed videos, including teams comfortable with self-hosting. It can also suit users who want configurable hosted or local models.
On the record
- Abstention
- A confidence gate can abstain without making an LLM call when neither retrieval branch clears its threshold.github.com · read 30 Sep 2026
- Citations
- Answers include clickable citations with matched keyframes, timestamps, transcript synchronization, evidence badges, and match scores.github.com · read 1 Oct 2026
- Default models
- The documented default stack uses OpenAI GPT-4o for answers, OpenAI text-embedding-3-small for transcripts, local CLIP for frames, fastembed for reranking, Whisper-1 for speech-to-text, and Gemini for speaker recognition.github.com · read 1 Oct 2026
- Deployment
- MomentSearch is packaged as a Docker image with api, worker, clip, and seed entrypoints and can deploy anywhere Docker runs.github.com · read 1 Oct 2026
- Input
- Users can upload videos or paste YouTube URLs for indexing.github.com · read 1 Oct 2026
- Integrations
- Documented infrastructure integrations include Postgres, Qdrant, Prefect, and object storage through Tigris, Amazon S3, or Google Cloud Storage.github.com · read 1 Oct 2026
- License
- MomentSearch is released under the Apache 2.0 license.github.com · read 1 Oct 2026
- Maker privacy
- Traversaal.ai says it uses strictly necessary authentication and usage cookies and does not use cookies for advertising, analytics, or tracking.traversaal.ai · read 30 Sep 2026
- Model flexibility
- Answer, embedding, reranking, and speech-to-text models are configurable by provider name, including hosted and local options.github.com · read 1 Oct 2026
- Models
- Answer, embedding, reranking, and speech-to-text models are pluggable by provider name, with OpenAI, Gemini, Anthropic, OpenRouter, xAI, Groq, Together, Fireworks, Mistral, NVIDIA, Azure OpenAI, Ollama, LM Studio, vLLM, Cohere, Voyage, Jina, and local CLIP options documented.github.com · read 30 Sep 2026
- Purpose
- MomentSearch lets users ask questions about videos and retrieves exact moments using what appears on screen and what is said in transcripts.github.com · read 1 Oct 2026
- Retrieval
- The system searches visual frames and transcripts in parallel, fuses the results, and reranks matching moments.github.com · read 1 Oct 2026
- Retrieval-only mode
- Without an LLM key, the application still returns ranked clickable moments but does not generate prose answers.github.com · read 1 Oct 2026
- Security limit
- The API is unauthenticated; anyone who can reach its port can mint upload URLs, ingest, ask, and delete, so the README recommends localhost binding or an authenticating proxy.github.com · read 30 Sep 2026
- Security limitation
- The API has no authentication, and anyone who can reach the port can upload, ingest, ask, and delete data.github.com · read 1 Oct 2026
- Speaker recognition
- Optional per-video speaker recognition uses Gemini video understanding to label who said each line.github.com · read 1 Oct 2026
- Storage integrations
- Object storage supports local storage, AWS S3, GCP storage, GCP native storage, and Fly Tigris, while the architecture uses Postgres, Qdrant, and Prefect Cloud.github.com · read 30 Sep 2026
- Storage protection
- The project recommends keeping buckets private and using presigned URLs for uploads and playback.github.com · read 1 Oct 2026
