mcpField
J
json-mender-mcp

json-mender-mcp

Local
github.com/headlessherm-creatorDeveloper Tools0Updated 1w ago

About json-mender-mcp

json-mender-mcp is a local MCP server for AI agents that repairs malformed JSON, validates JSON Schema, and tests regex patterns. Its tools run deterministically on the machine, with no API keys or network calls.

Key features

  1. 01Repair malformed JSON
  2. 02Validate data against JSON Schema
  3. 03Test regex patterns deterministically
  4. 04Run locally without network calls

Use cases

  1. 01Fix invalid JSON from LLM output
  2. 02Check data against a JSON Schema
  3. 03Verify regex matches target strings

MCP support

It exposes JSON repair, schema validation, and regex testing tools to MCP clients.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"headlessherm-creator-json-mender": {
"command": "npx",
"args": [
"-y",
"json-mender-mcp"
]
}
}
}
Run commandnpx -y json-mender-mcp

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmjson-mender-mcp0.2.1

Tools

3

Saved tool list · updated 1w ago

ParameterTypeRequiredDescription
textstringYesMalformed JSON text to repair.

Details

From the project README

JSON Mender MCP — Fix Malformed JSON, Validate JSON Schema, Test Regex for AI Agents

A free, local MCP server that fixes broken JSON output from LLMs, validates data against a real JSON Schema, and tests regex patterns deterministically — no API keys, no network calls, no telemetry.

If you're building an AI agent and keep hitting JSONDecodeError, unterminated strings, trailing commas, or unquoted keys in model output, this is the MCP server for that exact problem.

Same local-first, deterministic philosophy as polymath-megablaster-mcp, applied specifically to the "model wrote JSON that doesn't parse" problem.

The problem this solves

LLMs — especially smaller or local models (Llama, Qwen, Mistral) — routinely produce JSON that looks right but fails to parse: trailing commas, single quotes instead of double, unescaped inner quotes, missing commas, truncated output. Asking the model to "fix the JSON" usually just produces a second broken attempt. repair_json fixes it deterministically, every time, the same way, with no second model call needed.

Tools

  • repair_json — fix malformed/invalid JSON (trailing commas, unquoted keys, unescaped quotes, truncated structures) and return the corrected, parsed result. Built on jsonrepair.
  • validate_regex — test a regex pattern against one or more strings deterministically. Useful because models frequently write regex that compiles but doesn't match what they think it matches.
  • validate_against_schema — validate any JSON data against a real JSON Schema (via ajv, the most widely used JS Schema validator) and get precise, per-field, human-readable error messages instead of a generic pass/fail.

Setup

1. Install Node.js (v18+) if you don't have it: https://nodejs.org

2. Clone and install:

git clone https://github.com/headlessherm-creator/json-mender-mcp.git
cd json-mender-mcp
npm install

3. Point your MCP client at it:

{
  "mcpServers": {
    "json-mender": {
      "command": "node",
      "args": ["/path/to/json-mender-mcp/src/index.js"]
    }
  }
}

Setup docs for common clients:

Why local

No fetch/http import anywhere in src/tools.js — every tool runs as pure local computation using jsonrepair and ajv, both well-established, actively maintained npm packages. Nothing you send through this server leaves your machine.

Keywords

JSON repair MCP server, fix broken JSON from LLM, malformed JSON parser, JSON schema validation MCP, regex tester MCP, local MCP server no API key, Claude MCP JSON tools, AI agent structured output validation.

Related tools

Same local-first philosophy, different problems:

License

MIT

Comments

Markdown supported

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.