Obsidian Copilot

Executes dual-path search across the entire Obsidian vault using BM25+ lexical indexing as the default free tier, with optional embedding-backed vector search via Orama or Miyo APIs for semantic similarity. The indexing system maintains an in-memory inverted index of vault contents, while the retrieval layer implements RAG-style context envelope construction that ranks results by relevance and injects top-K documents into LLM prompts. Search results are ranked and formatted as markdown context blocks injected into chat messages.

Abstracts 15+ LLM providers (OpenAI, Anthropic, Google, Groq, Ollama, Azure, etc.) behind a unified ChatModelProviders enum and model management system. The chain execution system streams responses token-by-token from the selected provider's API, with built-in error handling and fallback logic. Supports both cloud-hosted APIs (via API keys) and local models (Ollama, LM Studio) without code changes, enabling users to swap providers without reconfiguring prompts or context handling.

The Plus-tier document parsing feature allows users to upload PDF, EPUB, and DOCX files, which are converted to markdown by Brevilabs' hosted backend and ingested into the vault. The conversion process extracts text, preserves structure (headings, lists, tables), and generates markdown files that can be searched and linked like native notes. This is a hosted service; documents are sent to Brevilabs' infrastructure for processing.

The Plus-tier 'Self-Host Mode' (Believer tier) allows users to replace Brevilabs' hosted backend with self-hosted services: Miyo for embeddings, Firecrawl for web scraping, and Perplexity for web search. This enables privacy-conscious deployments where all data remains under user control. Configuration is via settings UI, allowing users to point to their own instances of these services. The agent system automatically uses the configured backends for search and web access.

The settings UI allows users to configure multiple LLM providers (OpenAI, Anthropic, Google, etc.) with API keys, select default models for chat and embeddings, and customize behavior (context size, temperature, streaming, etc.). Settings are stored in Obsidian's plugin data directory and can be exported/imported. The interface supports both simple (API key + model) and advanced (custom endpoints, proxy settings) configuration. Model selection is dynamic; users can switch models without restarting Obsidian.

The plugin implements a standard Obsidian plugin lifecycle (onload, onunload) with lazy initialization of expensive components (embeddings, indexing, agent infrastructure). The state management system persists plugin state (settings, conversation history, memory notes) to Obsidian's plugin data directory, enabling recovery after crashes or restarts. The plugin integrates with Obsidian's command palette and ribbon UI for easy access to chat and commands.

Enables conversational chat with fine-grained control over which vault content is included in each message. Users can select specific notes, folders, or tags to inject as context, or use the free 'Vault QA' mode for full-vault search. The context envelope system constructs a ranked context block from selected sources, injecting it into the system prompt. The Plus tier 'Project Mode' allows defining scoped contexts from folders/tags/URLs, enabling multi-project workflows where different conversations operate over different knowledge domains.

Implements a ReAct (Reasoning + Acting) agent loop that enables the LLM to autonomously decide when to search the vault, fetch web content, or apply edits via the Composer tool. The agent maintains an internal reasoning trace, calls tools based on LLM-generated function calls, and iterates until reaching a terminal state (answer found, max steps exceeded, or error). Tools include vault search (BM25+/semantic), web search (via Firecrawl or Perplexity), and note editing (via Composer with diff preview). This is a Plus-tier feature backed by Brevilabs' hosted infrastructure.

The Composer tool allows users to request AI-generated edits to notes, with a diff preview interface showing proposed changes before application. Users can accept, reject, or manually edit the diff. The editing system integrates with Obsidian's file API to apply changes atomically, maintaining undo/redo history. Supports both inline edits (modify selected text) and full-note rewrites. The agent system can autonomously apply edits via Composer as part of tool-calling loops.

Allows users to define custom commands as markdown files with prompt templates, variable placeholders (e.g., {{selectedText}}, {{fileName}}), and optional YAML frontmatter for configuration. Commands are executed by substituting variables and sending the rendered prompt to the LLM. Supports both simple one-shot prompts and complex multi-step workflows. Commands are stored as markdown files in the vault, enabling version control and sharing. The command system integrates with the chat UI and can be triggered via slash commands or the command palette.

Allows users to attach images to chat messages and send them to vision-capable LLM models (GPT-4V, Claude 3, Gemini Vision, etc.). The chat UI handles image upload, encoding, and injection into multimodal prompts. Supports multiple image formats (PNG, JPEG, WebP) and can process images alongside text context from notes. The vision capability is provider-dependent; only models with vision support can process images.

The 'Relevant Notes' sidebar panel displays notes semantically related to the current chat or selected note, ranked by relevance. The panel combines semantic search results with link-graph analysis (backlinks, forward links, link distance) to surface both directly connected and semantically similar notes. Results are updated in real-time as the user types or selects notes. The panel integrates with the chat context system, allowing users to quickly inject suggested notes into the conversation.

Automatically saves chat conversations as markdown notes in a designated vault folder (configurable in settings). Each conversation is stored as a single markdown file with messages formatted as blockquotes or list items, preserving the full conversation history including context injections and tool calls. Conversations are searchable via Obsidian's native full-text search and can be linked to other notes. The system supports exporting conversations and integrates with Obsidian's sync and backup mechanisms.

The Plus-tier long-term memory feature allows agents to read and write persistent memory notes, enabling the agent to learn from previous interactions and maintain state across conversations. The agent can autonomously update memory notes with new information, insights, or summaries, and retrieve memory notes as context for future decisions. This is implemented via tool-calling, where the agent can invoke 'read memory' and 'write memory' tools. Memory notes are stored in the vault and can be manually edited by users.