Coming Soon
Features under active development
These features have interfaces defined in @localmode/core and are under active development. Provider implementations in @localmode/transformers and @localmode/webllm are in progress.
The features listed below are not yet production-ready. APIs may change before the stable release.
Classification & NLP
Text Classification
classify(), classifyMany() — Sentiment analysis, emotion detection, topic classification
Zero-Shot Classification
classifyZeroShot() — Classify text into arbitrary labels without fine-tuning
Named Entity Recognition
extractEntities(), extractEntitiesMany() — Extract people, organizations, locations, dates
Interfaces
import type { ClassificationModel, ZeroShotClassificationModel, NERModel } from '@localmode/core';Audio Processing
Speech-to-Text
transcribe() — Transcription with Whisper models, word-level timestamps
Text-to-Speech
synthesizeSpeech() — Generate speech audio from text
Interfaces
import type { SpeechToTextModel, TextToSpeechModel } from '@localmode/core';Vision
Image Classification
classifyImage(), classifyImageZeroShot() — Categorize images
Image Captioning
captionImage() — Generate descriptions with BLIP models
Object Detection
detectObjects() — Detect objects with bounding boxes
Image Segmentation
segmentImage() — Background removal, semantic segmentation
Image Features
extractImageFeatures() — Feature vectors for image similarity search
Image Transformation
imageToImage() — Super resolution, style transfer
Interfaces
import type {
ImageClassificationModel,
ZeroShotImageClassificationModel,
ImageCaptionModel,
ObjectDetectionModel,
SegmentationModel,
ImageFeatureModel,
ImageToImageModel,
} from '@localmode/core';Text Processing
Translation
translate() — Multi-language translation
Summarization
summarize() — Extractive and abstractive summarization
Fill-Mask
fillMask() — BERT-style masked token prediction
Question Answering
answerQuestion() — Extractive QA from context
Interfaces
import type {
TranslationModel,
SummarizationModel,
FillMaskModel,
QuestionAnsweringModel,
} from '@localmode/core';Document Understanding
OCR
extractText() — Optical character recognition from images
Document QA
askDocument() — Question answering on document images (invoices, forms)
Table QA
askTable() — Question answering on tabular data
Interfaces
import type { OCRModel, DocumentQAModel, TableQAModel } from '@localmode/core';Advanced RAG
BM25 Search
createBM25() — Keyword-based search for hybrid retrieval
Hybrid Search
hybridSearch(), reciprocalRankFusion() — Combine vector and keyword search