Jsonaut
@shoggi211
关于 Jsonaut
Repair malformed JSON from LLM tool calls and agent-to-agent handoffs — fixes trailing commas, single quotes, truncation, markdown code fences, comments, and python literals, and can validate/coerce the result against a JSON Schema. Free deterministic tier; pay-per-call LLM fallb
基本信息
配置
使用下面的配置,将此服务器添加到你的 MCP 客户端。
{
"mcpServers": {
"jsonaut": {
"type": "http",
"url": "https://jsonaut.jsonaut-shaurya.workers.dev/mcp"
}
}
}工具
未检测到工具
工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。
概览
What is Jsonaut?
Jsonaut is a tool that repairs malformed JSON produced by LLM tool calls and agent-to-agent handoffs. It fixes issues like trailing commas, single quotes, truncation, markdown code fences, comments, and Python literals, and can validate/coerce the result against a JSON Schema. It is built for developers working with LLM-generated JSON that needs to be clean.
How to use Jsonaut?
You can use Jsonaut via its MCP server (exposing repair_json and validate_json tools), its REST API, or by installing its local-first npm or Python client libraries. Most JSON fixes happen deterministically without external calls.
Key features of Jsonaut
- Repairs common JSON malformations from LLM output (trailing commas, single quotes, truncation, etc.)
- Validates and coerces results against a JSON Schema
- Free deterministic tier for most fixes
- Pay-per-call LLM fallback via x402 (USDC on Base)
- Exposes MCP endpoints (
repair_json,validate_json), REST API, and local npm/Python clients
Use cases of Jsonaut
- Cleaning malformed JSON from LLM tool call responses
- Fixing JSON during agent-to-agent handoffs
- Ensuring LLM-generated JSON conforms to a specified JSON Schema
- Adding a recovery layer in LLM pipelines without full retries
FAQ from Jsonaut
What types of JSON issues does Jsonaut fix?
It corrects trailing commas, single quotes, truncations, markdown code fences, comments, and Python literals in JSON output from LLMs.
Is there a free tier?
Yes, a free deterministic tier handles most fixes locally without any payment. A paid LLM fallback is available via x402 (USDC on Base) for edge cases.
How can I access Jsonaut?
Jsonaut is available as an MCP server, a REST API, and as local-first npm or Python client libraries.
Does Jsonaut require an internet connection?
The deterministic tier works offline with the local npm/Python clients. The LLM fallback (x402) requires an internet connection and a USDC payment on Base.
AI 与智能体 分类下的更多 MCP 服务器
Open Multi-Agent Canvas
CopilotKitThe open-source multi-agent chat interface that lets you manage multiple agents in one dynamic conversation and add MCP servers for deep research
meGPT - upload an author's content into an LLM
adriancoCode to process many kinds of content by an author into an MCP server
Just Prompt - A lightweight MCP server for LLM providers
dislerjust-prompt is an MCP server that provides a unified interface to top LLM providers (OpenAI, Anthropic, Google Gemini, Groq, DeepSeek, and Ollama)
Gemini MCP Server
aliargunMCP server implementation for Google's Gemini API
Sequential Thinking Multi-Agent System (MAS)
FradSerAn advanced sequential thinking process using a Multi-Agent System (MAS) built with the Agno framework and served via MCP.
评论