MCP Server

Ajoute la vérification de citations académiques à Claude Desktop, Cursor ou tout client compatible MCP.

@citetrue/mcp-servernpx -y @citetrue/mcp-server

Ce que vous obtenez

Deux outils qui enveloppent l'API REST CiteTrue. Mêmes crédits, même précision — simplement via MCP au lieu de HTTP.

Outil	But	Crédits / appel
verify	Split a text blob into references and verify each against real sources.	1 per ref; tricky ones up to 5
get_credits	Renvoie le solde de crédits du compte.	0

Démarrage rapide

Aller sur Tableau de bord → Clés API et cliquer sur Nouvelle clé API. Copier la clé commençant par sk_.
Ouvrir la config du client MCP (voir Configuration client) et ajouter le bloc citetrue.
Redémarrer le client MCP — pas juste le chat, toute l'application.
Demander à l'assistant "Vérifie ces citations : [coller]". Il appellera verify et résumera.

Configuration client

Tout client MCP stdio fonctionne. Les trois ci-dessous sont les plus courants ; les autres utilisent la même structure.

Claude Desktop

Recommandé

Installez CiteTrue comme extension Claude Desktop — sans édition de JSON ni installation de Node. Le serveur empaqueté s'exécute avec le Node fourni avec Claude Desktop.

Télécharger pour Claude Desktop

Double-cliquez sur le fichier téléchargé (ou glissez-le dans Paramètres → Extensions) et collez votre clé API lorsque c'est demandé — elle est stockée en sécurité dans le trousseau de votre système.

Important : après l'installation, assurez-vous que l'extension est activée dans Paramètres → Extensions, puis redémarrez Claude Desktop — c'est ce qui active ses outils.

Ou configurer manuellement

Éditer ~/Library/Application Support/Claude/claude_desktop_config.json sur macOS, %APPDATA%\Claude\claude_desktop_config.json sur Windows :

{
  "mcpServers": {
    "citetrue": {
      "command": "npx",
      "args": ["-y", "@citetrue/mcp-server"],
      "env": {
        "CITETRUE_API_KEY": "sk_..."
      }
    }
  }
}

Cursor

Dans Settings → MCP → Add new MCP server, utiliser la même structure command / args / env que le JSON Claude Desktop ci-dessous.

Windsurf, Zed, autres

Tout client MCP stdio fonctionne — quel que soit le chemin de config, les trois mêmes champs (command, args, env).

Configuration

Variable env	Par défaut	Notes
CITETRUE_API_KEY	— (Requis)	Token Bearer depuis le dashboard.
CITETRUE_API_URL	`https://api.citetrue.com`	Écraser pour des déploiements auto-hébergés.

Outils

verify

Split a free-form text blob into individual references and verify each against real sources. Every reference goes through the same flow, and each one in the result reports what it cost.

Paramètres

Supply exactly one of text or taskHash:

text string — raw text containing one or more references: numbered / bulleted / blank-line-separated lists, BibTeX, or prose with in-text author-year citations. Max 10MB.
taskHash string — resume a task started by an earlier call (the taskHash it returned). Finished results are replayed without charging again; citations that still need checking are completed and billed as usual.
force bool, optional — ignore cached results and check the text again from scratch (billed like a new check). Only applies with text.

Sortie

{
  "taskHash": "...",
  "refs": [
    {
      "hash": "a1b2c3...",
      "text": "[1] Vaswani, A., Shazeer, N., Parmar (2017)...",
      "assessment": "authentic",
      "value": {
        "type": "citation",
        "title": "Attention is all you need",
        "authors": ["A Vaswani", "N Shazeer", "N Parmar"],
        "year": 2017,
        "url": "https://..."
      },
      "note": "",
      "notices": [],
      "confidence": 0.97,
      "cost": 1
    }
  ],
  "cost": 2,
  "balance": 248
}

Each ref's cost is what the task has charged for that citation: 1 for most citations, up to 5 for tricky ones that needed more data sources and more analysis, 0 if it was already paid for in another task. Resuming a task with taskHash reports the same per-ref numbers again — they are not new charges. The top-level cost is what this call newly charged (0 when a finished task is simply replayed), and balance is what's left afterwards — it can be negative (see Billing). note is a free-form one-line natural-language explanation in English (intended for end-user display); for programmatic checks, match on notices[] instead.

Assessment values

authentic — paper exists and is the one cited; value contains the matched paper.
unsure — found something but can't confirm; notices and confidence describe why.
inauthentic — definitively not found / fabricated / unrelated.
invalid — input wasn't a citation; pipeline refused to search.
error — verification failed on this ref; errors[] lists tokens.
exceeded — credits ran out before this ref ran.

Facturation

Same credits as the REST API. Most citations cost 1 credit. Tricky ones that need more data sources and more analysis can cost up to 5. If you run out of credits mid-check, citations already in progress still finish; the overage is deducted at your next credit reset, so balance can be negative. Citations that couldn't start for lack of credits come back as exceeded. get_credits is free.

Dépannage

If the server fails to start, ensure npx is on PATH and CITETRUE_API_KEY is set.

Maintenu par l'équipe CiteTrue · Dernière mise à jour: 1 octobre 2026