Skip to main content
Glama
Code Reference Optimizer MCP Server

Code Reference Optimizer MCP Server

by FosterG4

Code Reference Optimizer MCP Server

Trust Score npm version node version license

An advanced MCP (Model Context Protocol) server that intelligently extracts minimal, relevant code context using AST parsing, analyzes code differences, and optimizes imports to dramatically reduce token usage for AI assistants.

Key Features

  • Smart Context Extraction: Uses AST parsing to identify and extract only relevant code sections

  • Multi-language Support: TypeScript/JavaScript, Python, Go, Rust, Java, C++, and more

  • Intelligent Caching: LRU cache with configurable persistence and customizable storage paths

  • Token Optimization: Filters unnecessary code while maintaining semantic completeness

  • Diff Analysis: Provides minimal, focused code differences with semantic understanding

  • Import Optimization: Eliminates unused imports and suggests consolidation opportunities

  • Configurable: Runtime configuration via tools with persistent settings

  • Simple Integration: stdio-based server and optional HTTP server, easy to integrate with any MCP client

Related MCP server: diffctx

Quick Start (STDIO)

  • One-off (recommended):

npx -y @fosterg4/mcpsaver

You should see: Code Reference Optimizer MCP server running on stdio.

  • Global (optional):

npm i -g @fosterg4/mcpsaver
mcpsaver

HTTP Mode

Start the HTTP server on port 8081 (default):

npx -y @fosterg4/mcpsaver mcpsaver-http
# or after build
npm run start:http

Configure your MCP client to use the HTTP binary if supported, e.g.:

{
  "mcpServers": {
    "mcpsaver": { "command": "mcpsaver-http", "env": { "PORT": "8081", "LOG_LEVEL": "info" } }
  }
}

Use with an MCP client

Add to your MCP client config (example mcpServers.json):

{
  "mcpServers": {
    "mcpsaver": {
      "command": "npx",
      "args": ["-y", "@fosterg4/mcpsaver"],
      "env": {}
    }
  }
}

Available Tools

The server provides 7 powerful tools for code analysis and optimization:

πŸ” extract_code_context

Extracts minimal, focused code context using AST parsing. Intelligently identifies relevant code sections, imports, and dependencies for specific symbols.

{
  "filePath": "path/to/file.ts",
  "targetSymbols": ["myFunc", "MyClass"],
  "includeImports": true,
  "maxTokens": 1000
}

Required: filePath | Optional: targetSymbols, includeImports, maxTokens

πŸ’Ύ get_cached_context

Retrieves previously extracted and cached code context for fast access without re-parsing.

{ "filePath": "path/to/file.ts", "cacheKey": "optional-key" }

Required: filePath | Optional: cacheKey

πŸ“Š analyze_code_diff

Performs intelligent analysis of code differences with semantic understanding and minimal update suggestions.

{
  "filePath": "path/to/file.ts",
  "oldContent": "export function a() { return 1 }",
  "newContent": "export function a() { return 2 }"
}

Required: filePath, oldContent, newContent

🧹 optimize_imports

Analyzes and optimizes import statements to eliminate redundancy and improve code efficiency.

{ "filePath": "path/to/file.ts", "usedSymbols": ["useEffect", "useMemo"] }

Required: filePath | Optional: usedSymbols

βš™οΈ get_config

Retrieves current configuration settings for cache behavior, extraction parameters, and more.

{ "section": "cache" }

Optional: section (cache, extraction, imports, diff, performance, languages, logging, security)

πŸ”§ update_config

Updates configuration settings including cache policies, token limits, and performance thresholds.

{
  "config": {
    "cache": { "enablePersistence": true, "persistencePath": "/custom/cache/path" },
    "extraction": { "maxTokens": 2000 }
  }
}

Required: config

πŸ”„ reset_config

Resets all configuration settings to default values.

{}

No parameters required

Note: Tool results are returned as MCP content with a single text item containing JSON of the result, e.g.

{
  "content": [{ "type": "text", "text": "{\n  \"...\": true\n}" }]
}

Additional MCP Capabilities

  • Prompts: listed but empty; get_prompt returns MethodNotFound.

  • Resources: listed but empty; reading a resource returns MethodNotFound.

  • Roots: exposes the current working directory as a single root workspace.

  • Sampling: stubbed; sampling/createMessage returns MethodNotFound.

Examples

See docs/EXAMPLES.md for end‑to‑end request examples of each tool.

Configuration

  • Call get_config, update_config, reset_config to manage runtime settings.

  • You may also set environment variables via your MCP client if supported (e.g., LOG_LEVEL).

Structured Logging

The server uses a lightweight structured logger. Configure via get_config/update_config or env:

  • Level: config.logging.level (trace|debug|info|warn|error)

  • File logging: config.logging.enableFileLogging and config.logging.logPath

Development

npm ci
npm run build
npm start     # run built server (stdio)
npm run start:http # run built server (http)
npm run dev   # tsc --watch
npm test
npm run lint
npm run type-check

Publishing (maintainers)

npm login
npm run clean && npm run build
npm version patch
npm publish --access public

License

MIT β€” see LICENSE.

Available Tools

7 tools
analyze_code_diffC

Analyze differences between code versions and provide minimal updates

Inputs: filePath (required), newContent (required), oldContent (required).

extract_code_contextC

Extract minimal code context from files using AST parsing

Inputs: filePath (required), includeImports, maxTokens, targetSymbols.

get_cached_contextB

Retrieve cached code context for a file

Inputs: filePath (required), cacheKey.

get_configC

Get current configuration settings

Inputs: section.

optimize_importsC

Analyze and optimize import statements to reduce redundancy

Inputs: filePath (required), usedSymbols.

reset_configB

Reset configuration to default values

Inputs: no named parameters.

update_configC

Update configuration settings

Inputs: config (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3.4/5.0

    Scored across 7 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose with no ambiguity: analyze_code_diff handles code version comparisons, extract_code_context and get_cached_context manage code context retrieval (one for extraction, one for caching), get_config, update_config, and reset_config handle configuration operations, and optimize_imports focuses on import optimization. The descriptions clearly differentiate their functions, making misselection unlikely.

    Naming Consistency5/5

    All tools follow a consistent verb_noun pattern using snake_case, such as analyze_code_diff, extract_code_context, and optimize_imports. This uniformity makes the tool set predictable and easy to understand, with no deviations in naming conventions.

    Tool Count5/5

    With 7 tools, the count is well-scoped for a code reference optimizer server. Each tool serves a specific role in code analysis, context management, and configuration, with no apparent bloat or missing essential functions, fitting the server's purpose appropriately.

    Completeness4/5

    The tool set covers core areas like code analysis, context extraction/caching, and configuration management, providing good lifecycle coverage. A minor gap exists in direct code editing or refactoring tools beyond imports, but agents can work around this using the provided tools for most optimization tasks.

    Maintenance

    ActivityInactive
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      C
      maintenance
      Provides intelligent code context and analysis through semantic compression, AST parsing, and multi-language support. Offers 60-80% token reduction while enabling AI assistants to understand codebases through local analysis, OpenAI-enhanced insights, and GitHub repository integration.
      6
      20 npm
      3
      MIT
    • A
      license
      A
      quality
      B
      maintenance
      Selects the minimum code context an LLM needs to understand a git diff: graph-based fragment selection under a token budget, deterministic output, 30+ tree-sitter languages.
      1
      315 PyPI
      6
      Apache 2.0
    • A
      license
      Not graded
      quality
      C
      maintenance
      Provides structured, token-efficient code context for AI assistants by indexing codebases with AST-level analysis, supporting 17+ languages, enabling semantic search, call-site tracking, and PR attribution.
      45 PyPI
      44
      MIT