capabilities.work

What you can do with NVIDIA NIM

Every task in plain language. Pick one, and we can help you get access.

22 things you can do

NVIDIA NIM

NVIDIA NIM (NVIDIA Inference Microservices) is a catalog of GPU-accelerated, containerized AI inference microservices that package optimized model engines (TensorRT-LLM, vLLM, SGLang, Triton) behind industry-standard OpenAI-compatible REST APIs. NIM covers large language models, embeddings and reranking, vision-language models, speech (Riva), visual generative AI, and biology (BioNeMo) — exposed i

7 multi-step recipes

Recipes: multi-step jobs

BioNeMo Drug Discovery
Fold a protein, dock a ligand, then generate analog molecules.Discover And Chat
Discover a served model and run a chat completion against it.Generate Image And Caption
Generate an image, then describe it with a VLM to produce a caption.Health Gated Completion
Gate a text completion behind the container readiness probe.RAG Rerank And Answer
Embed a query, rerank candidate passages, and answer grounded in the best one.Vision Describe And Summarize
Describe an image with a VLM, then summarize the description into a caption.Voice Assistant Loop
Speech-to-text, chat answer, then text-to-speech in a single loop.

Artificial Intelligence Management

Complete a raw text promptCreate embedding vectors for textTranscribe an audio clip to textGenerate or edit an image from a prompt

Model Inference

Predict a protein's 3D structure from its sequenceDock a small-molecule ligand to a proteinGenerate small molecules around a seed SMILES

Retrieval & Embeddings

Rerank passages by relevance to a query

Speech & Voice

Synthesize speech audio from text

Other things you can do

Generate a chat reply from a conversationList the models this NIM endpoint servesGet details for one modelCheck that the NIM container process is aliveCheck that the model is loaded and readyGet Prometheus metrics for a NIM container