Khoj

Indexes and searches across user's notes, documents, and web content using vector embeddings to retrieve contextually relevant information. Implements a unified search layer that abstracts over heterogeneous data sources (local files, cloud storage, web pages) and returns ranked results based on semantic similarity rather than keyword matching, enabling the agent to ground responses in user-specific context.

Generates natural language responses to user queries by combining retrieved context from the knowledge base with an underlying LLM (OpenAI, Anthropic, or local models). The system maintains conversation history, integrates retrieved documents into the prompt, and generates responses that cite specific sources, implementing a retrieval-augmented generation (RAG) pattern with explicit source attribution.

Maintains conversation history and context across multi-turn interactions, enabling the assistant to reference previous messages and maintain coherent dialogue. Implements context window management to fit conversation history and retrieved documents within LLM token limits, with strategies for summarization or selective context inclusion.

Allows users to configure LLM parameters (temperature, top-p, max tokens, etc.) and embedding model selection to tune assistant behavior and performance. Provides configuration interfaces for adjusting generation quality, response length, and semantic search sensitivity without code changes.

Provides a unified interface to multiple LLM providers (OpenAI, Anthropic, local/self-hosted models) allowing users to configure and switch between models without changing application code. Abstracts over provider-specific APIs and response formats, enabling model selection at runtime and supporting both cloud and local inference paths.

Extends the knowledge base with real-time web search capability, allowing the agent to retrieve current information from the internet when local documents don't contain relevant answers. Integrates web search results into the RAG pipeline, enabling responses grounded in both personal knowledge and current web content with source attribution for web pages.

Generates new content (articles, summaries, emails, code) by combining user prompts with relevant context from the knowledge base, enabling creation of documents grounded in personal information and style. Uses the underlying LLM with retrieved context to produce coherent, contextually-aware generated content that reflects the user's existing knowledge and preferences.

Enables users to define automated research and content tasks that run on a schedule or trigger, combining web search, knowledge base retrieval, and content generation into multi-step workflows. Supports task decomposition, progress tracking, and autonomous execution with human oversight, implementing a workflow orchestration layer on top of core capabilities.

Supports on-premises deployment where all data (documents, conversations, embeddings) remains local and never transmitted to cloud services. Enables users to run the full Khoj stack (search, generation, web integration) on their own infrastructure using local LLMs and embedding models, providing complete data privacy and control without cloud dependencies.

Provides native or web-based clients across multiple platforms (web, desktop, mobile) that connect to a central Khoj backend, maintaining synchronized conversation history and knowledge base access. Enables users to interact with the assistant from any device while maintaining consistent state and context across sessions.

Provides mechanisms for extending Khoj with custom integrations and plugins, allowing users to connect additional data sources, tools, and services. Supports integration with external APIs, document sources, and custom logic without modifying core Khoj code, enabling ecosystem expansion and customization for specific use cases.

Accepts and indexes documents in multiple formats (PDF, Markdown, plain text, Word documents, etc.) by extracting text content and converting to embeddings for semantic search. Handles document parsing, chunking, and metadata extraction to prepare content for the knowledge base, supporting both batch ingestion and incremental updates.