aidex_describe
Add or update a section in the project summary (summary.md). Use to document project purpose, architecture, key concepts, or patterns.
LABSCO SUMMARY
Two things at once — a cheaper way to search code than reading files, and a place for the assistant to keep what it learned.
Signatures and identifier lookup answer most "where is this defined, what does it take" questions without pulling whole files into the context, and semantic search covers the case where the name is unknown. The memory side is what changes across sessions: notes, tasks and a session start that reports what changed since last time, so a new chat does not begin from nothing. It stays on your machine unless you deliberately configure a model provider.
What it is
A local code-intelligence server: it builds a searchable index of a project's identifiers and signatures, keeps notes and tasks in its own database, and can receive live logs from a running program.
What you get
- An index of the project's identifiers, queried directly instead of grepping the tree, with incremental update or removal for a single file after an edit
- Signatures rather than whole files — header comments, types and method prototypes for one file or a glob of them, which is what a model needs to call something correctly
- Semantic search over embedded code, docs and workspace items for the times you know the concept but not the name
- Project orientation: an auto-detected summary of entry points, main types and languages, the indexed file tree with per-file statistics, and a browsable file listing
- Memory that outlives the chat — session notes, a session start that reports which files changed since last time, and a project summary you can write sections into
- A task backlog kept with the code: tasks created, read, updated and filtered by status
- Cross-project work: link a dependency's own index, scan a directory tree for indexed projects, register them globally, then query identifiers or signatures across all of them at once
- Persistent global guidelines stored as named values, so conventions apply across projects
- A log hub that receives output from any external program over HTTP with no SDK, so the assistant sees what the code actually did
- An interactive browser viewer for the project tree, with a settings tab for the embedding and LLM configuration
- Screenshots and window listing, for checking what an application is showing
Requirements
Node.js 20 or newer. Installed globally from npm as aidex-mcp, with a setup command that registers it with your assistant. Local-first: it works fully offline on its own embeddings, and the LLM layer — query translation and reranking — is opt-in, switched on by configuring an Anthropic, OpenAI, OpenRouter, Ollama or Hugging Face key. Embeddings are enabled per project when you initialize it and cached afterwards.
Setup effort
One command — npm install -g aidex-mcp
USE CASES
01
Getting a model oriented in an unfamiliar repository
Index the project, read the summary of entry points and main types, then walk the tree.
02
Finding where something is defined
Query the identifier index, then read the signatures for the files it points at.
03
Searching when you only know the concept
Use semantic search across code, docs and workspace items instead of guessing at names.
04
Picking up where the last session left off
Start the session to see what changed since last time, and read the notes you left.
05
Following a symbol into a dependency
Link the dependency's index, then query identifiers or method signatures across every registered project.
06
Seeing what the program actually did
Send the running application's logs to the hub over HTTP and query them alongside the code.
07
Keeping the index honest as you edit
Re-index the file you changed, and drop the ones you deleted.
31 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
aidex_files
List all files and directories in the indexed project. Returns the complete project structure with file types (code, config, doc, asset, test, other) and whether each file is indexed for code search. Use modified_since to find files changed in this session.
aidex_global_guideline
Manage persistent guidelines in the global AiDex database (~/.aidex/global.db). Guidelines are named key-value instructions that apply across all projects — e.g. "review" → detailed review checklist, "release-prep" → release steps, "new-feature" → conventions to follow. Use list at session start to load active guidelines. Works without prior global_init.
aidex_global_init
Scan a directory tree for AiDex-indexed projects and register them in the global database (~/.aidex/global.db). Also finds unindexed projects (by markers like .csproj, package.json, Cargo.toml, etc.). Use this to enable cross-project searches with aidex_global_query.
aidex_global_query
Search for a term across ALL registered projects in the global AiDex index. Returns matches grouped by project. Use this to find code across your entire codebase.
aidex_global_refresh
Refresh project statistics in the global AiDex index. Updates file counts, method counts, etc. from each project's database. Removes projects whose paths no longer exist.
aidex_global_signatures
Search for methods and types by name across ALL registered projects. Returns method prototypes and type definitions grouped by project. Use to find "where is class X defined?" or "who has a method called Y?".
aidex_global_status
Show overview of all projects registered in the global AiDex index. Lists project names, paths, file counts, languages, and last indexed times.
aidex_init
Initialize AiDex indexing for a project. Scans all source files and builds a searchable index of identifiers, methods, types, and signatures.
aidex_link
Link a dependency project to enable cross-project queries. The dependency must have its own .aidex index.
aidex_links
List all linked dependency projects.
aidex_log
Universal Log Hub — receive and query logs from any external program (C#, Python, Node, etc.) via HTTP. Zero-cost when not used. Actions: init (start HTTP server), free (stop server), status (show stats), query (search logs), clear (reset buffer), write (inject entry as "claude"), control_get (read all interactive dashboard control values), control_set (change one control — same set-point the user's dashboard slider drives, so the AI can tune live).
aidex_note
Read or write a session note for the project. Notes persist in the AiDex database. When a note is overwritten or cleared, the old note is automatically archived. Use history/search to browse past notes.
aidex_query
Search for terms/identifiers in the AiDex index. Returns file locations where the term appears. PREFERRED over Grep/Glob for code searches when .aidex/ exists - faster and more precise. Use this instead of grep for finding functions, classes, variables by name.
aidex_remove
Remove a file from the AiDex index. Use when a file has been deleted from the project.
aidex_scan
Scan a directory tree to find all projects with AiDex indexes (.aidex directories). Use this to discover which projects are already indexed before using Grep/Glob - indexed projects should use aidex_query instead.
aidex_screenshot
Take a screenshot. Returns file path for Read. No index required. OPTIMIZATION STRATEGY: Always start with scale=0.5 + colors=2 (smallest). If text is unreadable, retry with colors=16. If still unclear, try scale=0.75 or omit colors for full quality. Remember what works for each window/app during the session to avoid retries.
aidex_search
Semantic search across embedded code, docs, and workspace items (tasks/notes/history). Best for natural-language questions like "how do we handle retry with backoff" or "what does the error logging do" — finds the right file even when you don't know the identifier name. Requires embeddings: true on aidex_init for the project. Three modes: semantic (pure vector KNN), exact (identifier match like aidex_query), hybrid (RRF fusion of both — default and recommended).
aidex_session
Start or check an AiDex session. Call this at the beginning of a new chat session to: (1) detect files changed externally since last session, (2) auto-reindex modified files, (3) get session note and last session times. Returns info for "What did we do last session?" queries.
aidex_settings
Open the AiDex Settings tab in the viewer (where the user configures embeddings, LLM provider/key/model, and the privacy switch). Use this when the user asks to "open my AiDex settings", "show my LLM configuration", "configure embeddings", etc. Without 'open: true' the tool returns the current settings as JSON for inspection.
aidex_signature
Get the signature of a single file: header comments, types (classes/structs/interfaces), and method prototypes. Use this INSTEAD of reading entire files when you only need to know what methods/classes exist. Much faster than Read tool for understanding file structure.
aidex_signatures
Get signatures for multiple files at once using glob pattern or file list. Returns types and method prototypes. Use INSTEAD of reading multiple files when exploring codebase structure. Much more efficient than multiple Read calls.
aidex_status
Get AiDex server status and statistics for an indexed project
aidex_summary
Get project summary including auto-detected entry points, main types, and languages. Also returns content from summary.md if it exists.
aidex_task
Manage a single task in the project backlog. Actions: create (new task), read (get task + log), update (change fields), delete, log (add history note). Tasks persist in the AiDex database. Completed tasks are preserved as documentation.
aidex_tasks
List and filter tasks in the project backlog. Returns tasks grouped by status (active, backlog, done, cancelled) and sorted by priority. Use to get an overview of all open and completed work.
aidex_tree
Get the indexed file tree. Optionally filter by subdirectory, limit depth, or include statistics per file.
aidex_unlink
Remove a linked dependency project.
aidex_update
Re-index a single file. Use after editing a file to update the AiDex index. If the file is new, it will be added to the index. If unchanged (same hash), no update is performed.
aidex_viewer
Open an interactive project tree viewer in the browser. Shows the indexed file structure with clickable nodes - click on a file to see its signature (header comments, types, methods). Uses a local HTTP server with WebSocket for live updates.
aidex_windows
List all open windows with their titles, PIDs, and process names. Use this to find the exact window title for aidex_screenshot with mode="window". No project index required.
BEFORE YOU INSTALL
What to know first
Tools you get
31 tools when we connected.
Requires
Node.js ≥ 20
Where it runs
All 4 documented clients run the server on your machine.
CONTEXT COST
Moderate Standing cost
Rated by the size of the 31 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
TERMINAL
npm install -g aidex-mcpMCP Clients
Claude Code
CLAUDE CODE · JSON
{
"mcpServers": {
"aidex": {
"type": "stdio",
"command": "aidex",
"env": {}
}
}
}Claude Desktop, Gemini CLI
JSON
{
"mcpServers": {
"aidex": {
"command": "aidex"
}
}
}VS Code
VS CODE · JSON
{
"servers": {
"aidex": {
"type": "stdio",
"command": "aidex"
}
}
}
