Features

Mitii combines deep workspace indexing with a safe Plan/Act agent workflow.

Context and indexing

• Workspace scanner respects .gitignore and .thunderignore; auto-indexes on folder open
• FTS5 full-text search with ripgrep fallback for unindexed paths
• Symbol extraction for TypeScript, JavaScript, Python, Java, Go
• PageRank repo map — surfaces structurally important files
• Vector search — local MiniLM embeddings via @xenova/transformers; SQLite or LanceDB backend
• Hybrid retriever + reranker — merges search, vectors, mentioned files, git state, and LSP diagnostics

Agent workflow

• Plan / Act / Review modes — plan before you touch code; act with tools; review without rewriting
• Tool loop — read_file, search, write_file, apply_patch, run_command, git_diff, diagnostics, and more
• Research subagents — parallel read-only workers for exploration
• Task decomposition — multi-step plans with lifecycle tracking
• Post-edit verification — configurable lint/test commands after Act-mode runs
• Skills catalog — drop SKILL.md files in .thunder/skills/ and invoke via use_skill

Safety and control

• Approval modes — review_all, ask_edits, ask_deletes, ask_commands, auto
• Autonomy presets — safe, guided, builder, pilot, enterprise (network disabled)
• Untrusted workspace blocking — writes and shell disabled unless you opt in
• Auto-checkpoints before approved file writes
• Optional diff preview in VS Code before patches land

Memory and persistence

• Long-term memory — memory_search / memory_write with FTS5 + optional vector hybrid search
• Post-task memory extraction — observations captured after completed work
• Session history — resume from the History tab
• Plan persistence — plans saved to SQLite and .thunder/tasks/

MCP and integrations

Preloaded keyless MCP servers (disable with thunder.mcp.preloadBuiltin: false):

Server	Purpose
filesystem	Scoped file access for the open workspace
memory	Cross-session knowledge graph
sequential-thinking	Structured reasoning helper

Add your own via thunder.mcp.servers, .thunder/mcp.json, or .mcp.json.

UI

React sidebar webview with chat, history, settings, approval cards, subagent activity panel, plan panel, checkpoint browser, indexing status, context budget warnings, and token meter.

How Mitii compares

Pain point	What Mitii does
Agent doesn't know the codebase	Background workspace index with FTS, symbols, vectors, repo map
Wrong files in context	Hybrid retrieval + reranker + token budgeter
Edits without oversight	Approval queue for writes and shell; tunable autonomy presets
Plans that never get executed	Plan mode persists steps; Act mode runs the tool loop
Context runs out mid-task	Conversation compaction, auto-continue, task state between pauses
No audit trail	JSONL session logs in .thunder/logs/
Locked into one vendor's rules	Loads AGENTS.md, .cursor/rules, .clinerules, and .thunder/rules