Connectez LLMtoMD à vos outils d'IA
Un seul serveur MCP, tous les assistants. Exploitez vos documents — convertir, rechercher et interroger — dans Claude, ChatGPT, Cursor et vos propres agents. Connectez-vous pour les apps ; une clé API pour le code.
La mémoire de votre agent de code
Les agents de code oublient. À mesure que la conversation s'allonge, la spécification sort du contexte et l'IA se met à deviner. Conservez vos exigences dans LLMtoMD et votre agent pourra les retrouver — le détail exact, à la demande — au lieu de perdre le fil.
Importez votre FRD
Déposez votre document d'exigences, votre spécification ou vos notes de conception — n'importe quel fichier ou lien.
Nous le structurons
LLMtoMD le convertit en Markdown clair et interrogeable dans votre bibliothèque.
Votre agent se souvient
Cursor, Claude Code, VS Code ou Antigravity l'interrogent comme un contexte durable.

Claude
Exploitez vos documents dans Claude.ai, Claude Desktop et la Claude API.
Connector URL
https://mcp.llmtomd.com/mcpClaude.ai (web)
- Open claude.ai → Settings → Connectors → Add custom connector (a paid plan is required).
- Paste the Connector URL above, leave the OAuth fields blank, and click Connect.
- Sign in to LLMtoMD and click Allow.
Claude Desktop
- Settings → Developer → Edit config, add this to claude_desktop_config.json, then restart. The first run opens a browser to sign in:
{
"mcpServers": {
"llmtomd": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://mcp.llmtomd.com/mcp"]
}
}
}Claude API
- Pass LLMtoMD as a remote MCP server with an API key:
client.beta.messages.create(
model="claude-opus-4-8",
max_tokens=1024,
mcp_servers=[{
"type": "url",
"name": "llmtomd",
"url": "https://mcp.llmtomd.com/mcp",
"authorization_token": "mic_YOUR_KEY",
}],
betas=["mcp-client-2025-11-20"],
messages=[{"role": "user", "content": "List my LLMtoMD documents."}],
)mic_YOUR_KEY par une clé issue de Paramètres → Connectez vos outils d'IA.Obtenir une clé API 
Cursor
Donnez vos documents à l'IA de Cursor pendant que vous codez.
Connector URL
https://mcp.llmtomd.com/mcp- Click Add to Cursor above and confirm the install — no file editing.
- Cursor opens a browser to sign in to LLMtoMD the first time — no key needed.
- Prefer to do it by hand? Create ~/.cursor/mcp.json (global) or .cursor/mcp.json (project) with the entry below instead.
{
"mcpServers": {
"llmtomd": { "url": "https://mcp.llmtomd.com/mcp" }
}
}
VS Code
Offrez à votre agent de code VS Code une base de connaissances persistante — FRDs, spécifications et documents.
Connector URL
https://mcp.llmtomd.com/mcpGitHub Copilot (agent mode)
- Click Add to VS Code above, then Install in the prompt VS Code shows — no file editing.
- The first run opens a browser to sign in to LLMtoMD (OAuth) — no key. After you approve, the tools appear in Copilot's agent mode.
- Prefer to do it by hand? Command Palette → MCP: Open User Configuration and add the server below instead.
{
"servers": {
"llmtomd": { "type": "http", "url": "https://mcp.llmtomd.com/mcp" }
}
}Claude Code (in VS Code or the terminal)
- Claude Code has no one-click link (it keeps its own config). Run this once, then /mcp to finish the sign-in:
claude mcp add --transport http llmtomd https://mcp.llmtomd.com/mcp
Claude Code — share with a repo (.mcp.json)
- Commit a .mcp.json at the repo root so everyone who opens it gets LLMtoMD (Claude Code prompts to approve on first use):
{
"mcpServers": {
"llmtomd": { "type": "http", "url": "https://mcp.llmtomd.com/mcp" }
}
}
Antigravity
Laissez l'agent de Google Antigravity consulter vos spécifications et documents de conception pendant qu'il développe.
- Antigravity Settings → Customizations → Open MCP Config (or edit ~/.gemini/config/mcp_config.json). Note: Antigravity uses serverUrl, not url.
- Generate an API key below and paste it into the Authorization header (recommended — Antigravity's sign-in for custom servers is new and can hang mid-OAuth; the key always works).
- Save and restart Antigravity; the LLMtoMD tools appear under Manage MCP servers.
{
"mcpServers": {
"llmtomd": {
"serverUrl": "https://mcp.llmtomd.com/mcp",
"headers": { "Authorization": "Bearer mic_YOUR_KEY" }
}
}
}mic_YOUR_KEY par une clé issue de Paramètres → Connectez vos outils d'IA.Obtenir une clé API 
ChatGPT
Ajoutez LLMtoMD comme connecteur dans ChatGPT.
Connector URL
https://mcp.llmtomd.com/mcp- Enable Developer Mode: ChatGPT → Settings → Connectors (Apps) → Advanced → Developer mode (Plus/Pro/Team/Enterprise, beta).
- In a chat, open the + (or Settings → Connectors) → Add a custom connector with the Connector URL above.
- Sign in to LLMtoMD and click Allow.

OpenAI API
Appelez LLMtoMD depuis l'OpenAI Responses API ou l'Agents SDK.
Responses API
from openai import OpenAI
client = OpenAI()
resp = client.responses.create(
model="gpt-5",
tools=[{
"type": "mcp",
"server_label": "llmtomd",
"server_url": "https://mcp.llmtomd.com/mcp",
"authorization": "mic_YOUR_KEY",
"require_approval": "never",
}],
input="List my LLMtoMD documents.",
)
print(resp.output_text)Agents SDK
from agents import Agent, Runner
from agents.mcp import MCPServerStreamableHttp
async with MCPServerStreamableHttp(
name="LLMtoMD",
params={
"url": "https://mcp.llmtomd.com/mcp",
"headers": {"Authorization": "Bearer mic_YOUR_KEY"},
},
) as server:
agent = Agent(name="Assistant", instructions="Use the LLMtoMD tools.", mcp_servers=[server])
print((await Runner.run(agent, "List my LLMtoMD documents.")).final_output)mic_YOUR_KEY par une clé issue de Paramètres → Connectez vos outils d'IA.Obtenir une clé API 
LangChain
Chargez les outils de LLMtoMD dans un agent LangChain / LangGraph.
LangSmith / Fleet (no code)
- In LangSmith → Settings → MCP servers → Add server → Add custom MCP.
- Name it "llmtomd" and set the URL to https://mcp.llmtomd.com/mcp.
- Under Authentication keep Static Headers, click Add header, and set Key to Authorization and Value to the line below. (Or choose OAuth 2.1 (Auto) to sign in instead of pasting a key.)
- Save server — the LLMtoMD tools now show under Tools and are available to your Fleet agents.
Authorization: Bearer mic_YOUR_KEY
Python (langchain-mcp-adapters)
- Install the adapter: pip install langchain-mcp-adapters
from langchain_mcp_adapters.client import MultiServerMCPClient
client = MultiServerMCPClient({
"llmtomd": {
"transport": "streamable_http",
"url": "https://mcp.llmtomd.com/mcp",
"headers": {"Authorization": "Bearer mic_YOUR_KEY"},
}
})
tools = await client.get_tools() # pass these into your agentmic_YOUR_KEY par une clé issue de Paramètres → Connectez vos outils d'IA.Obtenir une clé API 
LlamaIndex
Utilisez les outils de LLMtoMD dans un agent LlamaIndex.
- Install the tools package: pip install llama-index-tools-mcp
from llama_index.tools.mcp import BasicMCPClient, McpToolSpec
client = BasicMCPClient("https://mcp.llmtomd.com/mcp", headers={"Authorization": "Bearer mic_YOUR_KEY"})
tools = McpToolSpec(client=client).to_tool_list() # pass these into your agentmic_YOUR_KEY par une clé issue de Paramètres → Connectez vos outils d'IA.Obtenir une clé API Exploitez vos documents dans chaque assistant
Convertissez, recherchez et interrogez vos fichiers — dans Claude, ChatGPT, Cursor et vos propres agents — via un seul serveur MCP.