LocalMode

Privacy-first AI chat with streaming responses and vision support. All processing happens locally.

AI agent that uses tools to research topics step-by-step. Watch reasoning unfold in real-time.

Inspect GGUF model metadata, check browser compatibility, and chat via WASM. Access 160K+ models from HuggingFace.

Ask questions about your PDF documents with semantic search and reranking.

Build your personal knowledge base with semantic and hybrid search.

Full LangChain RAG pipeline running locally. Document ingestion, semantic search, and LLM answers with source citations.

Import vectors from Pinecone, ChromaDB, CSV, or JSONL. Export to interoperable formats. Re-embed text-only records locally.

Record audio and transcribe with Moonshine. Search your voice notes semantically.

Transcribe meetings, generate summaries, extract action items.

Convert text to natural speech with Kokoro TTS. 29 English voices, streaming playback.

Explore and compare Kokoro TTS voices. Browse all 29 voices, synthesize text, compare side-by-side.

Analyze sentiment of reviews and feedback. Batch process thousands of texts.

Classify emails by intent with custom labels. Route to folders automatically.

Translate text between 20+ languages. Works completely offline.

Summarize long documents into key points. Extractive and abstractive modes.

Ask questions about documents and get precise answers with confidence scores.

Intelligent text completion using fill-mask models. Writing assistant.

Ask questions about invoices, receipts, and forms. Visual document understanding.

AI writing assistant with summarization and translation. Instant with Chrome AI, automatic fallback to Transformers.js.

Extract structured JSON from text using local LLMs. Schema-validated, retry with self-correction.

Detect device capabilities, get ranked model recommendations, and compute optimal batch sizes. No model download required.

Evaluate classification models with accuracy, precision, recall, F1, and confusion matrix. Calibrate similarity thresholds from corpus embeddings.

Remove backgrounds from images automatically. Download as PNG with transparency.

Auto-categorize photos, search by description, find similar images.

Search products by image, find similar items, auto-categorize catalog.

Generate alt-text for images automatically. Ask questions about images.

Extract text from images and scanned documents. Supports tables, formulas, and 11 languages.

Detect and locate objects in images. Real-time webcam support.

Upscale and enhance images with super resolution. Restore old photos.

Find visually similar and duplicate images. Clean up your photo library.

Search photos by text description or reference image. CLIP embeds text and images in the same vector space.

Real-time hand, pose, and face tracking, gesture recognition, audio classification, and language detection — all on-device with Google MediaPipe.

Detect and redact PII (names, locations, organizations) from documents.

End-to-end encrypted notes and documents. Privacy-first storage.