Examples catalog (116)

Indexed from modal-examples. Use the slug to pin an example on an event.

Slug	Title	Themes	GPU	Run
`generators`	Run a generator function on Modal			`modal run 01_getting_started/generators.py`
`hello_world`	Hello, world!	getting_started, batch		`modal run 01_getting_started/hello_world.py`
`install_cuda`	Installing the CUDA Toolkit on Modal	web	T4	`modal run 02_building_containers/install_cuda.py`
`install_flash_attn`	Install Flash Attention on Modal	vision	L40S	`modal run 02_building_containers/install_flash_attn.py`
`basic_grid_search`	Hyperparameter search	rag, vision, getting_started		`modal run 03_scaling_out/basic_grid_search.py`
`cls_with_options`	Override Modal resource options (GPU, scaling) at runtime with `Cls.with_options`		T4	`modal run 03_scaling_out/cls_with_options.py`
`dynamic_batching`	Dynamic batching for ASCII and character conversion	batch		`modal run 03_scaling_out/dynamic_batching.py`
`db_to_sheet`	Write to Google Sheets from Postgres	web		`modal deploy 04_secrets/db_to_sheet.py`
`hackernews_alerts`	Run cron jobs in the cloud to search Hacker News	rag, vision		`modal run 05_scheduling/hackernews_alerts.py`
`schedule_simple`	Scheduling remote jobs	getting_started		`modal run 05_scheduling/schedule_simple.py`
`whisperx_transcribe`	WhisperX transcription with word-level timestamps	speech, vision	H100	`modal run 06_gpu_and_ml/audio-to-text/whisperx_transcribe.py`
`esmfold2_binder_design`	Design protein binders at scale with ESMFold2 and ESMC		H100	`modal run 06_gpu_and_ml/binder-design/esmfold2_binder_design.py`
`blender_video`	Render a video with Blender on many GPUs or CPUs in parallel	vision, batch	L40S	`modal run 06_gpu_and_ml/blender/blender_video.py`
`controlnet_gradio_demos`	Play with the ControlNet demos	vision, web	A10G	`modal run 06_gpu_and_ml/controlnet/controlnet_gradio_demos.py`
`amazon_embeddings`	Embed 30 million Amazon reviews at 575k tokens per second with Qwen2-7B	rag, batch, llm	L40S	`modal run 06_gpu_and_ml/embeddings/amazon_embeddings.py`
`image_embeddings_infinity`	Modal Cookbook: Recipe for Inference Throughput Maximization	rag, batch, vision	gpu	`modal run 06_gpu_and_ml/embeddings/image_embeddings_infinity.py`
`liquidai_embeddings_server`	Serve Liquid AI ColBERT embeddings with llama.cpp and Modal Servers	rag, batch, llm		`modal run 06_gpu_and_ml/embeddings/liquidai_embeddings_server.py`
`gpu_fallbacks`	Set "fallback" GPUs		subprocess	`modal run 06_gpu_and_ml/gpu_fallbacks.py`
`gpu_snapshot`	Snapshot GPU memory to speed up cold starts		a10	`modal deploy 06_gpu_and_ml/gpu_snapshot.py`
`hp_sweep_gpt`	Train an SLM from scratch with early-stopping grid search over hyperparameters	llm, vision, rag	A10G	`modal run 06_gpu_and_ml/hyperparameter-sweep/hp_sweep_gpt.py`
`image_to_video`	Animate images with Lightricks LTX-Video via CLI, API, and web UI	vision, web	H100	`modal run 06_gpu_and_ml/image-to-video/image_to_video.py`
`potus_speech_qanda`	Retrieval-augmented generation (RAG) for question-answering with LangChain	rag, llm, speech		`modal run 06_gpu_and_ml/langchains/potus_speech_qanda.py`
`chat_with_pdf_vision`	Chat with PDF: RAG with ColQwen2	llm, rag, vision	A100-80GB	`modal run 06_gpu_and_ml/llm-serving/chat_with_pdf_vision.py`
`deepseek_v4`	Serve DeepSeek V4 Pro on Modal with SGLang	llm, vision	f	`modal run 06_gpu_and_ml/llm-serving/deepseek_v4.py`
`deepseek_v4_flash`	Deploy DeepSeek-V4-Flash with SGLang and Modal	llm	f	`modal run 06_gpu_and_ml/llm-serving/deepseek_v4_flash.py`
`gpt_oss_inference`	Run OpenAI's gpt-oss model with vLLM	llm	f	`modal run 06_gpu_and_ml/llm-serving/gpt_oss_inference.py`
`inkling_small`	Serve Inkling-Small on Modal with SGLang	llm, vision	f	`modal run 06_gpu_and_ml/llm-serving/inkling_small.py`
`lfm_snapshot`	Low Latency, Serverless LFM2 with vLLM and Modal	llm, rag	GPU	`modal run 06_gpu_and_ml/llm-serving/lfm_snapshot.py`
`ministral3_inference`	Serverless Ministral 3 with vLLM and Modal	llm, vision	f	`modal run 06_gpu_and_ml/llm-serving/ministral3_inference.py`
`nemotron_inference`	Low latency Nvidia Nemotron 3 with SGLang and Modal	llm	GPU	`modal run 06_gpu_and_ml/llm-serving/nemotron_inference.py`
`sglang_low_latency`	Low latency Qwen 3.6 with SGLang and Modal	llm	GPU	`modal run 06_gpu_and_ml/llm-serving/sglang_low_latency.py`
`sglang_snapshot`	Serverless Qwen 3-8B with SGLang and Modal Snapshots	llm	GPU	`modal run 06_gpu_and_ml/llm-serving/sglang_snapshot.py`
`sglang_vlm`	Serve the Qwen 3.6 Vision-Language Model with SGLang	llm, vision, web, getting_started	GPU	`modal run 06_gpu_and_ml/llm-serving/sglang_vlm.py`
`stepfun_inference`	Run StepFun models with SGLang	llm	GPU	`modal run 06_gpu_and_ml/llm-serving/stepfun_inference.py`
`trtllm_latency`	Serve an interactive language model app with low-latency TensorRT-LLM (LLaMA 3 8B)	llm	GPU	`modal run 06_gpu_and_ml/llm-serving/trtllm_latency.py`
`trtllm_throughput`	Serverless TensorRT-LLM (LLaMA 3 8B)	llm	GPU	`modal run 06_gpu_and_ml/llm-serving/trtllm_throughput.py`
`very_large_models`	Serve very large language models (DeepSeek V3, Kimi-K2, GLM 4.7/5)	llm, rag	f	`modal run 06_gpu_and_ml/llm-serving/very_large_models.py`
`vllm_inference`	Run OpenAI-compatible LLM inference with Gemma and vLLM	llm, web	f	`modal run 06_gpu_and_ml/llm-serving/vllm_inference.py`
`vllm_low_latency`	Low latency Qwen 3 8B with vLLM and Modal	llm, getting_started	GPU	`modal run 06_gpu_and_ml/llm-serving/vllm_low_latency.py`
`vllm_throughput`	Run LLM inference at maximum throughput	llm, getting_started	GPU	`modal run 06_gpu_and_ml/llm-serving/vllm_throughput.py`
`long-training`	Run long, resumable training jobs on Modal	training	a10g	`modal run 06_gpu_and_ml/long-training.py`
`fine_tune_asr`	Fine-tune Whisper to Improve Transcription on Domain-Specific Vocab	speech, training	H100	`modal run 06_gpu_and_ml/openai_whisper/fine_tune_asr.py`
`paint_flowers`	Painting flowers with code	training, llm, vision		`modal run 06_gpu_and_ml/paint_flowers/paint_flowers.py`
`boltz_predict`	Fold proteins with Boltz-2		H100	`modal run 06_gpu_and_ml/protein-folding/boltz_predict.py`
`chai1`	Fold proteins with Chai-1	getting_started	H100	`modal run 06_gpu_and_ml/protein-folding/chai1.py`
`esm3`	Build a protein folding dashboard with ESM3, Molstar, and Gradio	vision, web	A10G	`modal run 06_gpu_and_ml/protein-folding/esm3.py`
`esmfold2`	Fold proteins and biomolecular complexes with ESMFold2		H100	`modal run 06_gpu_and_ml/protein-folding/esmfold2.py`
`isaac_lab`	Train a robot in a simulated environment with Isaac Lab	training	f	`modal run 06_gpu_and_ml/reinforcement-learning/isaac_lab.py`
`mjlab_motion_imitation`	Teach a humanoid to imitate motion with mjlab	training	A10	`modal run 06_gpu_and_ml/reinforcement-learning/mjlab_motion_imitation.py`
`segment_anything`	Run Facebook's Segment Anything Model 2 (SAM 2) on Modal	vision	A100	`modal run 06_gpu_and_ml/sam/segment_anything.py`
`batched_whisper`	Fast Whisper inference using dynamic batching	speech, batch	a10g	`modal run 06_gpu_and_ml/speech-to-text/batched_whisper.py`
`parakeet_multitalker`	Parakeet Multi-talker Speech-to-Text	speech		`modal run 06_gpu_and_ml/speech-to-text/parakeet_multitalker.py`
`sortformer2_1_speaker_diarization`	Streaming Speaker Diarization with Sortformer2.1	speech	L4	`modal run 06_gpu_and_ml/speech-to-text/sortformer2_1_speaker_diarization.py`
`streaming_kyutai_stt`	Stream transcriptions with Kyutai STT	speech	a10g	`modal deploy 06_gpu_and_ml/speech-to-text/streaming_kyutai_stt.py`
`streaming_parakeet`	Streaming audio transcription using Parakeet	speech	a10g	`modal run 06_gpu_and_ml/speech-to-text/streaming_parakeet.py`
`flux`	Run Flux fast on H100s with `torch.compile`	vision	H100	`modal run 06_gpu_and_ml/stable_diffusion/flux.py`
`image_to_image`	Edit images with Flux Kontext	vision, getting_started	B200	`modal run 06_gpu_and_ml/stable_diffusion/image_to_image.py`
`text_to_image`	Run Stable Diffusion 3.5 Large Turbo as a CLI, API, and web UI	vision, web, batch	H100	`modal run 06_gpu_and_ml/stable_diffusion/text_to_image.py`
`swe_gym`	Train coding agents	agents, sandboxes, training, web		`modal run 06_gpu_and_ml/swe_gym/swe_gym.py`
`tensorflow_tutorial`	TensorFlow tutorial	vision	T4	`modal run 06_gpu_and_ml/tensorflow/tensorflow_tutorial.py`
`chatterbox_tts`	Create a Chatterbox TTS API on Modal	speech, llm, web	a10g	`modal run 06_gpu_and_ml/text-to-audio/chatterbox_tts.py`
`generate_music`	Make music with ACE-Step 1.5	llm, vision	l40s	`modal run 06_gpu_and_ml/text-to-audio/generate_music.py`
`ltx`	Generate videos from prompts with Lightricks LTX-Video	vision	H100	`modal run 06_gpu_and_ml/text-to-video/ltx.py`
`ltx2_two_stage`	High-quality text-to-video with LTX-2	vision, training	H200	`modal run 06_gpu_and_ml/text-to-video/ltx2_two_stage.py`
`mochi`	Text-to-video generation with Mochi	vision	H100	`modal run 06_gpu_and_ml/text-to-video/mochi.py`
`torch_profiling`	Tracing and profiling GPU-accelerated PyTorch programs on Modal			`modal run 06_gpu_and_ml/torch_profiling.py`
`text_to_world`	Generate 3D worlds from text with LTX-2.3 and InSpatio-World	vision, training	H200	`modal run 06_gpu_and_ml/world-models/text_to_world.py`
`finetune_yolo`	Fine-tune open source YOLO models for object detection	vision, training, batch	TRAIN	`modal run 06_gpu_and_ml/yolo/finetune_yolo.py`
`badges`	Serve a dynamic SVG badge	vision, getting_started		`modal run 07_web/badges.py`
`basic_web`	Hello world wide web!	getting_started	h100	`modal run 07_web/basic_web.py`
`count_faces`	Run OpenCV face detection on an image	vision		`modal run 07_web/count_faces.py`
`fastapi_app`	Deploy FastAPI app with Modal	web		`modal run 07_web/fastapi_app.py`
`fastrtc_flip_webcam`	Run a FastRTC app on Modal			`modal deploy 07_web/fastrtc_flip_webcam.py`
`flask_app`	Deploy Flask app with Modal	web		`modal run 07_web/flask_app.py`
`flask_streaming`	Deploy Flask app with streaming results with Modal			`modal run 07_web/flask_streaming.py`
`server`	Deploy HTTP Servers with ultra low latency on Modal	llm		`modal run 07_web/server.py`
`server_affinity`	Affinity routing for Modal Servers	llm, getting_started		`modal run 07_web/server_affinity.py`
`sessioned_chatroom`	Build a sessioned chat room	llm		`modal run 07_web/sessioned_chatroom.py`
`streaming`	Deploy a FastAPI app with streaming responses	web		`modal run 07_web/streaming.py`
`webrtc_yolo`	Real-time object detection with WebRTC and YOLO	web, vision	A100-40GB	`modal run 07_web/webrtc/webrtc_yolo.py`
`generators_async`	Run async generator function on Modal			`modal run 08_advanced/generators_async.py`
`hello_world_async`	Async functions	getting_started		`modal run 08_advanced/hello_world_async.py`
`parallel_execution`	Parallel execution on Modal with `spawn` and `gather`	batch		`modal run 08_advanced/parallel_execution.py`
`poll_delayed_result`	Polling for a delayed result on Modal			`modal run 08_advanced/poll_delayed_result.py`
`dicts_and_queues`	Use Modal Dicts and Queues together			`modal run 09_job_queues/dicts_and_queues.py`
`doc_ocr_jobs`	Run a job queue that turns documents into structured data with Datalab Marker	vision	l40s	`modal run 09_job_queues/doc_ocr_jobs.py`
`doc_ocr_webapp`	Serve a receipt parsing web app	web, vision		`modal run 09_job_queues/doc_ocr_webapp.py`
`pipeline_orchestration`	Orchestrate a multi-step pipeline with Modal Functions			`modal deploy 09_job_queues/pipeline_orchestration.py`
`web_job_queue_wrapper`	Create a web wrapper for job queue, submission, polling, & results	llm, vision, web, getting_started		`modal run 09_job_queues/web_job_queue_wrapper.py`
`algolia_indexer`	Algolia docsearch crawler	rag		`modal deploy 10_integrations/algolia_indexer.py`
`cron_datasette`	Publish interactive datasets with Datasette	batch		`modal deploy 10_integrations/cron_datasette.py`
`dbt_duckdb`	Build your own data warehouse with DuckDB, DBT, and Modal	rag		`modal deploy 10_integrations/dbt/dbt_duckdb.py`
`mcp_server_stateless`	Deploy a remote, stateless MCP server on Modal with FastMCP	agents		`modal run 10_integrations/mcp_server_stateless.py`
`multion_news_agent`	MultiOn: Twitter News Agent	agents, web, vision		`modal run 10_integrations/multion_news_agent.py`
`pushgateway`	Publish custom metrics with Prometheus Pushgateway	web		`modal deploy 10_integrations/pushgateway.py`
`s3_bucket_mount`	Analyze NYC yellow taxi data with DuckDB on Parquet files from S3			`modal run 10_integrations/s3_bucket_mount.py`
`serve_streamlit`	Run and share Streamlit apps	web		`modal deploy 10_integrations/streamlit/serve_streamlit.py`
`modal_tailscale`	Add Modal Apps to Tailscale			`modal run 10_integrations/tailscale/modal_tailscale.py`
`webscraper`	A simple web scraper			`modal deploy 10_integrations/webscraper.py`
`webscraper_old`	Web Scraping on Modal	web		`modal run 10_integrations/webscraper_old.py`
`agent`	Build a coding agent with Modal Sandboxes and LangGraph	agents, llm, sandboxes	T4	`modal run 13_sandboxes/codelangchain/agent.py`
`langserve`	Deploy LangChain and LangGraph applications with LangServe	agents, web, sandboxes, llm		`modal run 13_sandboxes/codelangchain/langserve.py`
`computer_use_vnc`	Watch a Browser Use agent drive Chromium over VNC	agents, llm, sandboxes, web		`modal run 13_sandboxes/cua/computer_use_vnc.py`
`harbor_evals`	Run massively parallel evals on Modal Sandboxes	sandboxes, batch, llm, agents		`modal run 13_sandboxes/harbor_evals.py`
`jupyter_sandbox`	Run a Jupyter notebook in a Modal Sandbox	sandboxes, vision	None	`modal run 13_sandboxes/jupyter_sandbox.py`
`safe_code_execution`	Run arbitrary code in a sandboxed environment	sandboxes, vision		`modal run 13_sandboxes/safe_code_execution.py`
`sandbox_agent`	Run Claude Code in a Modal Sandbox	sandboxes, agents, vision, getting_started		`modal run 13_sandboxes/sandbox_agent.py`
`sandbox_pool`	Maintain a pool of warm Sandboxes that are healthy and ready to serve requests	sandboxes		`modal run 13_sandboxes/sandbox_pool.py`
`sandbox_secret_injection`	Inject secrets into Sandbox HTTPS requests with an Outbound Policy	sandboxes, web, agents		`modal run 13_sandboxes/sandbox_secret_injection.py`
`sidecar_agent`	Separate an agent from its execution environment with Sidecars	agents, sandboxes		`modal run 13_sandboxes/sidecar_agent.py`
`sidecar_redis`	Test a web app with Redis in a Sidecar	sandboxes, vision		`modal run 13_sandboxes/sidecar_redis.py`
`sidecar_secrets_injection`	Inject Secrets into a Sandbox with a proxy Sidecar	sandboxes, web		`modal run 13_sandboxes/sidecar_secrets_injection.py`
`sidecar_traffic_routing`	Filter Sandbox HTTPS traffic with a proxy Sidecar	sandboxes, web		`modal run 13_sandboxes/sidecar_traffic_routing.py`
`simple_code_interpreter`	Build a stateful, sandboxed code interpreter	sandboxes		`modal run 13_sandboxes/simple_code_interpreter.py`
`miles_grpo`	Make LLMs better at math	training, llm, cluster	f	`modal run 14_clusters/miles_grpo.py`
`simple_torch_cluster`	Simple PyTorch cluster	cluster, getting_started, training	GPU	`modal run 14_clusters/simple_torch_cluster.py`