MCP.so
ログイン

Summarization Functions

@Braffolk

Summarization Functions について

Provides summarised output from various actions that could otherwise eat up tokens and cause crashes for AI agents

基本情報

カテゴリ

その他

トランスポート

stdio

公開者

Braffolk

設定

標準の設定はありません

このサーバーの README には解析可能な MCP 設定ブロックが含まれていません。インストール手順はリポジトリをご確認ください。

リポジトリ

ツール

ツールは検出されませんでした

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is Summarization Functions?

An MCP server that provides intelligent text summarization for AI agents, primarily developed to enhance context window management and prevent overflow. It integrates with the Model Context Protocol and supports multiple AI providers (Anthropic, OpenAI, OpenAI-compatible, Google).

How to use Summarization Functions?

Install via npm i mcp-summarization-functions, then add the server to your MCP configuration file with required environment variables (PROVIDER and API_KEY). Use tools like summarize_command, summarize_files, summarize_directory, summarize_text, and get_full_content.

Key features of Summarization Functions

  • Command output summarization with execution.
  • File content analysis with technical accuracy.
  • Directory structure overview with recursion.
  • Flexible model support across providers.
  • AI agent context optimization with caching.

Use cases of Summarization Functions

  • Summarizing large command outputs to prevent AI agent context overflow.
  • Analyzing file contents focused on security, API surface, or errors.
  • Getting quick overviews of complex directory structures.
  • Summarizing arbitrary text like logs, errors, or API responses.

FAQ from Summarization Functions

What AI providers are supported?

Anthropic (Claude), OpenAI (GPT), OpenAI‑compatible (e.g., Azure), and Google (Gemini).

How is it used with AI agents?

Include the provided “CONTEXT MANAGEMENT” prompt in agent instructions to mandate summarization for all large outputs (commands, files, directories, etc.).

What tools does the server expose?

summarize_command, summarize_files, summarize_directory, summarize_text, and get_full_content.

What environment variables are required?

PROVIDER (e.g., ANTHROPIC) and API_KEY. Optional: MODEL_ID, PROVIDER_BASE_URL, MAX_TOKENS, SUMMARIZATION_CHAR_THRESHOLD, SUMMARIZATION_CACHE_MAX_AGE, and MCP_WORKING_DIR.

Is there a caching mechanism?

Yes; summarized content is cached with a default maximum age of 1 hour (SUMMARIZATION_CACHE_MAX_AGE). Full content is stored for later retrieval via get_full_content.

コメント

「その他」の他のコンテンツ