# TokenJar MCP server

> Reduces token consumption for AI coding assistants by up to 95% using AST skeletons, differential caching, and intelligent code slicing.

- Repository: https://github.com/Farukes/TokenJar
- Page: https://hostedmcp.org/servers/farukes/tokenjar
- Transport: —
- Stars: 6

## Tools (10)

- `find_symbol_global`
- `find_symbol_references`
- `tool_get_code_skeleton`
- `tool_get_symbol`
- `read_file_smart`
- `run_command_smart`
- `filter_output`
- `get_repo_map_tool`
- `get_directory_tree_tool`
- `cache_stats`