MCP SERVER · OFFICIAL PROJECTCheck what a codebase costs in tokens before you feed it to a model, find the files driving that number, and pull the whole thing as paginated markdown.
LABSCO SUMMARY
The standard failure is asking for a repository, getting most of a context window back, and discovering afterwards that a lockfile and a vendored dependency accounted for most of it. Reporting the token count first, and then which files are responsible, turns that into a decision made in advance. Pagination on the retrieval is the matching piece — a merged codebase that exceeds the response limit is not partially useful, it fails. Three tools, no writes, and the discipline is the product rather than the packaging.
A three-tool codebase packaging server built around token cost: measure it first, find the largest contributors, then retrieve the merged content in pages.
Nothing to supply — it reads the codebase it is pointed at.
One command — npx -y context-coder --mini --stdio
USE CASES
Get the token count before processing anything, so the decision is made on a number rather than after a failed call.
Find the largest files driving the count — usually lockfiles, vendored code or generated output — and exclude them deliberately.
Retrieve the merged markdown in pages, so a large repository arrives usably rather than truncated.
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
get_codebaseGenerate a a merged markdown file of the entire codebase.Results are paginated.If more content exists, a message will prompt to call again with the next page number. Leave the _pageSize variable empty UNLESS you are running under Claude Code, in that case set it to 30000 to work around the Claude Code limits.
get_codebase_sizeCheck the codebase size and token counts before processing. Returns token counts for Claude and ChatGPT, warns if the codebase is too large, and shows the largest files. IMPORTANT: You should ALWAYS run this tool at the start of EVERY NEW CONVERSATION before any other operations. After running this tool, you should then call get_codebase to retrieve the actual code.
get_codebase_top_largest_filesReturns the top X largest files in the codebase. Use this tool when users want to see more large files beyond the initial 10 shown in get_codebase_size. Helpful for identifying which files to add to .cocoignore for large codebases.
BEFORE YOU INSTALL
3 tools when we connected; the project documents 12. The install command runs mini mode (--mini): get_codebase_size, get_codebase and get_codebase_top_largest_files. The default full mode gives 11, and --edit-file-mode adds edit_file for 12.
npx on PATH
Both documented clients run the server on your machine.
INSTALLATION
npx -y context-coder --mini --stdio{
"mcpServers": {
"context-coder": {
"command": "npx",
"args": [
"-y",
"supergateway",
"--streamableHttp",
"http://localhost:3001/mcp"
]
}
}
}{
"mcpServers": {
"context-coder": {
"command": "npx",
"args": [
"-y",
"context-coder",
"--mini",
"--stdio"
]
}
}
}MCP servers
fkiene
yamadashy
raullenchai
yzfly