About configdoc
configdoc scans supplied source text for direct environment-variable reads and produces a sorted inventory with defaults, read counts, and file locations. It is for teams documenting configuration surfaces; its scan is local and does not read files or use the network.
Key features
- 01Extracts process.env and import.meta.env reads
- 02Captures literal defaults after ?? or ||
- 03Counts reads and lists source files
- 04Validates inputs before producing results
- 05Offers CLI and MCP server
Use cases
- 01Document environment options across source files
- 02Check which files read each option
- 03Review configuration defaults and usage counts
MCP support
The MCP server provides an extract_config_options tool. Clients pass source files as path and content objects and receive the configuration inventory.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
1Saved tool list · updated 1w ago
| Parameter | Type | Required | Description |
|---|---|---|---|
| files | array | Yes | Source files to scan (at least one) |
Details
From the project READMEconfigdoc
Every config option in the code, with its default, in one document. You pass configdoc the source files as text, and it scans them for direct environment reads — the config surface most code bases actually use — and returns a single inventory:
- Both read forms —
process.env.NAMEandimport.meta.env.NAMEare extracted, where NAME follows the uppercase env convention ([A-Z_][A-Z0-9_]*). Lowercase or mixed-case names are not matched, and destructured reads (const { X } = process.env) and non-env config objects are out of scope by design. - Defaults — when a read is immediately followed by
??or||and a literal, that literal is the read's default: a string literal contributes the raw text between its quotes (escapes are not interpreted), a number literal contributes itself as written. A fallback that is an identifier, a call, or anything else captures no default. - One entry per option — each distinct name appears
once, sorted ascending, with the default captured by the
first read of that name (in scan order: files in input
order, lines top to bottom, matches left to right) that
captures one —
nullwhen no read does — plus the total number of reads of that name across all files and the distinct files it is read in, sorted ascending.
Input is validated strictly: a missing or empty file set, a
file missing path or content, an empty, non-string, or
duplicate path, or a non-string content is an error naming
the offending path (for example files[0].content must be a string), and no partial inventory is produced. The scan is
pure and local: no network, no configuration, no files read
beyond the input you pass in.
CLI
npx @reneturcios/configdoc extract --input args.json
npx @reneturcios/configdoc --version
args.json contains the full arguments object:
{
"files": [
{
"path": "src/config.ts",
"content": "const port = process.env.PORT ?? 8080;\nconst host = process.env.HOST ?? \"localhost\";"
},
{
"path": "web/env.ts",
"content": "export const api = import.meta.env.API_URL;"
}
]
}
The inventory is printed as JSON:
{
"options": [
{ "name": "API_URL", "default": null, "reads": 1, "files": ["web/env.ts"] },
{ "name": "HOST", "default": "localhost", "reads": 1, "files": ["src/config.ts"] },
{ "name": "PORT", "default": "8080", "reads": 1, "files": ["src/config.ts"] }
],
"total_options": 3,
"files_scanned": 2
}
Exit codes: 0 on success, 1 on a runtime error
(including an unreadable or invalid input file, or invalid
input), 2 on a usage error.
MCP server
npx -y @reneturcios/configdoc mcp
MCP client config:
{
"mcpServers": {
"configdoc": {
"command": "npx",
"args": ["-y", "@reneturcios/configdoc", "mcp"]
}
}
}
Tools
-
extract_config_options— Extract every configuration option read in the code, with the default each read falls back to, in one inventory.files(array, required, at least one) — source files as{path, content}objects;pathis unique across the set andcontentis the full source text (it may be empty)
Returns
{options, total_options, files_scanned}, where each option is{name, default, reads, files}.
Docker
docker run -i ghcr.io/cloudcorprecords/configdoc
docker run --rm -v "$PWD:/w" ghcr.io/cloudcorprecords/configdoc extract --input /w/args.json
Development
npm install
npm run build # tsc -> dist/
npm test # tsx --test test/*.test.ts (pure local, no network)
npm run dev # run the CLI from source
License
MIT — Copyright (c) 2026 Rene Turcios
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.