"Optimizing AI Model Thinking, Token Usage, and Context Size" matching MCP servers:

Provides owner-bound secure storage and retrieval of tool outputs via MCP tools, enabling trusted persistence for agents like Claude Code, Codex CLI, and OpenCode.

MCP server that reduces token consumption in AI coding assistants by up to 90% via structural reads, PreToolUse hooks, and tp-\* subagents.

Enables Claude Desktop to retrieve only the most relevant sections from locally indexed documents, reducing context-window and token usage during conversations. Indexing and search run entirely on the user's machine using SQLite FTS5, with no network connection or telemetry.

Enables agents to download single model weight files from Hugging Face repositories or Civitai model versions directly into allowlisted ComfyUI model folders. Source identifiers, destination categories, and file extensions are validated, and existing files are never overwritten.

MCP server that helps AI coding agents understand a repository by providing lightweight tree/map, code search, and token-budgeted context packing tools without dumping the entire monorepo into the prompt.

Enables AI coding agents to persist structured long-term memory (gotchas, architecture, API notes) in a .context folder and sync across devices and agents via Git.

MCP server providing lossless file reads with deduplication, per-repo token metering, and HMAC-signed context receipts. Enables auditable, vendor-neutral measurement of what an AI agent saw.

Provides LLMs with secure, read-only access to local documentation by scanning directories, extracting content from PDF, DOCX, Markdown, and text files, and performing keyword searches.

A unified context layer that connects your local data — repositories, documents, remote machines, and notes — to LLM interfaces through the Model Context Protocol (MCP).

Enables local, bounded routing and evidence judgments for Codex tasks, with tools for reading file context, checking outputs, and selecting registered workflows while keeping execution and review with the client.

Provides AI agents with secure, read-only file system access to analyze and understand project codebases, enabling multi-repository context aggregation and cross-project code tracing.

A Model Context Protocol (MCP) server that provides VSCode context and filesystem operations for AI assistants.

This server provides an API to query Large Language Models using context from local files, supporting various models and file types for context-aware responses.

Enables ChatGPT to inspect a selected local project on demand, retrieving project overviews, code search results, file excerpts, and Git context without pasting entire repositories into the conversation.

A thin, local, read-only MCP gateway that exposes governed context documents and producer manifests as logical resources, enabling AI clients to access them without arbitrary filesystem access.

MCP server that gives AI chat real vision by reading images, files, logs, and zip archives via local OCR in any language, with privacy-first processing.

Compress OCR-heavy PDFs into dense packed images so agents can work with long visual documents.

Enables managing personal information with dynamic topic-based organization (tasks, meetings, contacts, etc.), supporting optional OTP authentication and AES-256 encryption for sensitive data with automatic backups.

Enables AI agents to search whole NTFS volumes by filename without any background GUI, parsing the $MFT directly and refreshing results from the USN Journal on demand. Supports an Everything-compatible query subset (boolean, path:, ext:, size:, date and regex filters), plus index status and full rebuild tools.

A simple MCP server for local documentation with RAG capabilities, enabling AI assistants to access and search local documents.

Enables Codex to gather bounded, read-only source evidence through literal search, optional file inventory, and exact line-range reads instead of automatically receiving entire files.

TaskFlow AI is a premium MCP Server designed for AI editors, offering 40+ built-in tools covering file system, HTTP requests, databases, Git operations, and more.

Provides a standardized way to interact with MinIO object storage, allowing access to text files, binary files, and bucket contents while supporting operations like listing buckets/objects, retrieving objects, and uploading files.

Provides a bulk_read_files tool that reads specified files and forwards them to a remote backend for summarization, so agents receive concise summaries instead of raw file contents.

Local-first MCP server for safely searching, reading, summarizing, tagging, deduplicating, and organizing local files with scoped access, read-only defaults, and dry-run plans.

Enables managing Linux hosts and local coding agents from an MCP client through shell commands, files, background jobs, system information, instruction and skill discovery, and optional Claude Code, Codex, or Cursor CLI conversations. It provides privileged remote access with token authentication and can run longer work in the background.

Enables MCP clients to inspect local Git repositories through five focused, read-only tools for repository trees, literal text search, numbered file reads, bounded git diffs, and project summaries while confining paths, skipping secret-like and binary files, and enforcing output limits.

Enables Claude on Windows to use local programs, files, and AI models through a localhost-only MCP server, with plugins for transcription, Office/PDF processing, study-card generation, folder watching, and UI automation. It can also generate, test, and enable new plugins on demand.

MCP server providing filesystem allowlist and pack-aware Qdrant RAG search. Enables reading files, listing directories, and performing filtered RAG searches with case and pack IDs.

Enables AI agents to read, search, create, and append Markdown notes from a local vault (e.g., Obsidian) via MCP, allowing natural-language note management through compatible clients.