Hugging Face · Capability

Hugging Face Model Inference

Unified workflow for running AI/ML inference across Hugging Face APIs, combining the Inference API, Inference Providers, and Text Generation Inference for NLP, vision, audio, and multimodal tasks. Used by ML engineers and AI application developers.

Run with Naftiko Hugging FaceMachine LearningInferenceAIText Generation

What You Can Do

POST
Run inference — Run inference on a model via the Inference API
/v1/inference/{model_id}
POST
Text generation — Generate text using the Inference API
/v1/text-generation/{model_id}
POST
Chat completion — Create chat completion via providers
/v1/chat/completions
POST
Text completion — Create text completion via providers
/v1/completions
POST
Create embeddings — Create text embeddings
/v1/embeddings
POST
Generate image — Generate images from text
/v1/images/generations
POST
Classify text — Classify text into categories
/v1/classification/{model_id}
POST
Summarize — Summarize text content
/v1/summarization/{model_id}
POST
Translate — Translate text between languages
/v1/translation/{model_id}

MCP Tools

run-inference

Run inference on any Hugging Face model by model ID.

read-only
generate-text

Generate text using a language model via the Inference API.

read-only
classify-text

Classify text into predefined categories.

read-only
answer-question

Answer questions based on provided context.

read-only
summarize-text

Summarize text content.

read-only
translate-text

Translate text between languages.

read-only
fill-mask

Fill in masked tokens in text.

read-only
extract-features

Extract feature vectors from text for embeddings.

read-only
classify-image

Classify images into categories.

read-only
detect-objects

Detect objects in images.

read-only
transcribe-speech

Transcribe audio to text using automatic speech recognition.

read-only
generate-image

Generate images from text prompts.

read-only
zero-shot-classify

Classify text without predefined training labels.

read-only
compute-similarity

Compute similarity between sentences.

read-only
providers-chat-completion

Create chat completion via OpenAI-compatible multi-provider API.

read-only
providers-text-completion

Create text completion via multi-provider API.

read-only
providers-create-embeddings

Create text embeddings via multi-provider API.

read-only
providers-generate-image

Generate images via multi-provider API.

read-only
providers-transcribe

Transcribe audio via multi-provider API.

read-only
providers-text-to-speech

Convert text to speech via multi-provider API.

read-only
tgi-generate

Generate text using the TGI native endpoint.

read-only
tgi-chat-completions

Create chat completions using TGI OpenAI-compatible Messages API.

read-only
tgi-tokenize

Tokenize input text and return token IDs.

read-only
tgi-server-info

Get TGI server information and deployed model details.

read-only
list-provider-models

List models available across all inference providers.

read-only

APIs Used

hf-inference hf-providers hf-tgi