---
title: "Examples by task"
description: "Existing Pixeltable notebooks, recipes, how-tos, and docs examples grouped by task: video search, PDF RAG, image dedup, audio search, and agent memory."
url: "https://www.pixeltable.com/examples"
updated: "2026-10-28"
---

# Examples by task

Notebooks, /recipes apps, /learn/how-to guides, /use-cases, /reference-architecture pages, and docs.pixeltable.com cookbooks, grouped by the task they teach.

Tutorials stay on [how-to guides](https://www.pixeltable.com/learn/how-to). Runnable apps stay on [example apps](https://www.pixeltable.com/recipes). Notebooks stay on [notebooks](https://www.pixeltable.com/notebooks). Architectures stay on [reference architectures](https://www.pixeltable.com/reference-architecture).

## Video search {#video-search}

Sample stills from a clip, embed them, optionally transcribe speech, and return a title plus a presentation timestamp.

- [Video search example app](https://www.pixeltable.com/recipes/video-search) (recipe) - Runnable app.py: frames, CLIP, HTTP search.
- [How do I index a video library and search it by text?](https://www.pixeltable.com/learn/how-to/search-a-video-library-by-text) (how-to) - TableModel walkthrough with FastAPIRouter.
- [How do I extract keyframes from video?](https://www.pixeltable.com/learn/how-to/extract-keyframes-from-video) (how-to) - frame_iterator, fps, keyframes_only, frame_attrs.
- [How do I build video RAG in Python?](https://www.pixeltable.com/learn/how-to/build-video-rag) (how-to) - Retrieved moment, then an LLM column.
- [How does video search work?](https://www.pixeltable.com/learn/how-video-search-works) (learn) - Concept: a moment, not a filename.
- [Video search engine use case](https://www.pixeltable.com/use-cases/video-search-engine) (use-case) - Implementation guide on the use-cases hub.
- [Object detection in videos notebook](https://www.pixeltable.com/notebooks/object-detection-in-videos) (notebook) - YOLOX on sampled frames (Colab).
- [Video search reference architecture](https://www.pixeltable.com/reference-architecture/video-search) (architecture) - Slots, Twelve Labs / Mixpeek alternatives, sketch.
- [Pixeltable for media and entertainment](https://www.pixeltable.com/industries/media-and-entertainment) (industry) - Worked trailer-and-archive pipeline.
- [Extract frames from videos (docs)](https://docs.pixeltable.com/howto/cookbooks/video/video-extract-frames) (docs) - docs.pixeltable.com cookbook.

## PDF RAG {#pdf-rag}

Split PDFs into passage rows, embed them, retrieve, then generate from those rows. Vision RAG is the image half of the same schema.

- [RAG demo notebook](https://www.pixeltable.com/notebooks/rag-demo) (notebook) - Chunking, embeddings, and LLM integration in Colab.
- [How do I build a multimodal RAG pipeline?](https://www.pixeltable.com/learn/how-to/build-a-multimodal-rag-pipeline) (how-to) - document_splitter plus CLIP; retrieve then generate.
- [How do I chunk PDFs for RAG?](https://www.pixeltable.com/learn/how-to/chunk-pdfs-for-rag) (how-to) - sentence,token_limit vs page metadata.
- [What is multimodal RAG?](https://www.pixeltable.com/learn/what-is-multimodal-rag) (learn) - Passages, stills, or frames as retrieved pieces.
- [Production RAG implementation](https://www.pixeltable.com/use-cases/production-rag-implementation) (use-case) - Use-case guide for a production RAG table.
- [Invoice audit example app](https://www.pixeltable.com/recipes/invoice-audit) (recipe) - Document rows and computed checks in app.py.
- [Multimodal RAG reference architecture](https://www.pixeltable.com/reference-architecture/multimodal-rag) (architecture) - Qdrant, pgvector, S3 slots and a TableModel sketch.
- [Pixeltable for document-heavy insurance](https://www.pixeltable.com/industries/document-heavy-insurance) (industry) - FNOL passages filtered by claim_id.
- [Build a RAG pipeline (docs)](https://docs.pixeltable.com/howto/cookbooks/agents/pattern-rag-pipeline) (docs) - docs.pixeltable.com cookbook.

## Image dedup {#image-dedup}

Hash exact duplicates, CLIP for near-duplicates and text-to-image search, check a candidate before insert.

- [How do I deduplicate a large image dataset?](https://www.pixeltable.com/learn/how-to/deduplicate-images-with-embeddings) (how-to) - Hash UDF, CLIP, keep_if_new. POST does not check.
- [How do I embed images for search?](https://www.pixeltable.com/learn/how-to/embed-images-for-search) (how-to) - CLIP EmbeddingIndex; similarity(string=) or similarity(image=).
- [How do I serve image similarity search?](https://www.pixeltable.com/learn/how-to/serve-image-similarity-search) (how-to) - FastAPIRouter insert plus text and image-upload search.
- [Embedding indexes notebook](https://www.pixeltable.com/notebooks/embedding-indexes) (notebook) - Incremental indexes and similarity search.
- [Travel scrapbook example app](https://www.pixeltable.com/recipes/travel-scrapbook) (recipe) - Image uploads in a runnable app.py.
- [User-uploaded media architecture](https://www.pixeltable.com/reference-architecture/user-uploaded-media-pipeline) (architecture) - S3, Qdrant, hash, CLIP, HTTP slots.
- [Pixeltable for ecommerce catalogs](https://www.pixeltable.com/industries/ecommerce-catalogs) (industry) - SKU photos with exact-hash and CLIP search.
- [Find similar images with CLIP (docs)](https://docs.pixeltable.com/howto/cookbooks/search/search-similar-images) (docs) - docs.pixeltable.com cookbook.

## Audio transcription and search {#audio-transcription-and-search}

Split uploads into timestamped slices, transcribe, embed the text, and return a file plus seconds.

- [How do I transcribe audio uploads and make them searchable?](https://www.pixeltable.com/learn/how-to/transcribe-audio-and-search-it) (how-to) - audio_splitter, Whisper, EmbeddingIndex, HTTP.
- [Audio processing and transcription notebook](https://www.pixeltable.com/notebooks/audio-transcriptions) (notebook) - Extract, transcribe, and search audio from video.
- [Audio transcription pipeline use case](https://www.pixeltable.com/use-cases/audio-transcription-pipeline) (use-case) - Implementation guide on the use-cases hub.
- [Working with OpenAI notebook](https://www.pixeltable.com/notebooks/working-with-openai) (notebook) - Includes Whisper among other OpenAI models.
- [Transcribe audio (docs)](https://docs.pixeltable.com/howto/cookbooks/audio/audio-transcribe) (docs) - docs.pixeltable.com cookbook.

## Agent memory {#agent-memory}

Durable memory rows, an embedding index, recall as a query, then a prompt. The loop is separate from the store.

- [Research Agent Memory example app](https://www.pixeltable.com/recipes/research-agent-memory) (recipe) - Runnable memory table in app.py.
- [Chat agent example app](https://www.pixeltable.com/recipes/chat-agent) (recipe) - A chat agent on the same backend.
- [Agent memory backend architecture](https://www.pixeltable.com/reference-architecture/agent-memory-backend) (architecture) - Mem0, Letta, Zep, Qdrant slots and a TableModel sketch.
- [AI agent architecture (interactive diagram)](https://www.pixeltable.com/ai-agent-architecture) (architecture) - Broader agent system diagram already on the site.
- [AI agents and MCP use case](https://www.pixeltable.com/use-cases/ai-agents-mcp) (use-case) - Agents that call Pixeltable as a tool.
- [What is agentic RAG?](https://www.pixeltable.com/learn/what-is-agentic-rag) (learn) - Retrieval an agent calls against durable rows.
- [Build an agent with memory (docs)](https://docs.pixeltable.com/howto/cookbooks/agents/pattern-agent-memory) (docs) - docs.pixeltable.com cookbook.