Program leads
Localization program oversight
Spot lagging markets and glossary drift — no CSV exports.
“Which language is furthest behind?”
New · Included with your plan
ConveyThis Serveur MCP pour Claude Code, Cursor et VS Code
Ask about translation progress, glossary terms, untranslated pages, and usage — in plain language, from the tools you already use.
Included on your plan Revocable keys Under a minute to connect
Everything your AI assistant can help with
Translation answers where you already work
- Project & language status
- Translation & glossary lookup
- Works in Cursor & Claude Code
- Usage statistics & team insights
What is Model Context Protocol (MCP)?
Model Context Protocol (MCP) is an open standard that lets your AI assistant pull live data from external tools — instead of you copying exports into a chat window.
ConveyThis implements MCP for translation workflows: projects, languages, strings, glossary terms, rules, and usage — all accessible from Cursor, Claude Code, and VS Code. See the setup guide or announcement.
Pourquoi utiliser le serveur MCP ConveyThis ?
Running multilingual sites means constant questions — German progress, glossary updates, untranslated URLs before launch. You shouldn’t have to log in and export data every time.
ConveyThis MCP brings those answers into Cursor, Claude Code, and VS Code. Same production data, formatted for your AI assistant — no custom scripts. Included on your plan; pair with WordPress, Shopify, or our REST API when you need deploy-time automation.
Comment fonctionne ConveyThis MCP
Connect your AI assistant to https://mcp.conveythis.com/mcp with a bearer token — it returns live translation context so you can ask in plain language.
- 1
Open MCP / AI Agents in your dashboard. Copy the key — shown once.
- 2
Ajouter ConveyThis à votre client
Same endpoint and bearer token for Cursor, Claude Code, VS Code, and more. Per-client configs.
- 3
Ask your first question
“How far along is German?” “Which site is behind?” Live answers from your production translation data.
Example prompts & workflows
Ask in plain language from Cursor, Claude Code, or VS Code — your AI assistant returns live data from your account.
ConveyThis MCP
Live answers in your editor
How far along is my Spanish translation?
Spanish is at 87% — 24 pages still need review, mostly product variants.
Which translations are outdated?
14 French pages changed after the last update — mostly product descriptions from this week.
Find every page where we translated 'checkout'.
12 URLs across /cart, /checkout, and 3 collection templates.
What's our approved German term for 'dashboard'?
Glossary rule: “Dashboard” → “Übersicht” (locked term, used on 8 pages).
Which pages have missing translations?
Spanish is missing /pricing, /about, and 3 collection pages added after your last Shopify sync.
How many words have I used this month?
128,400 words — 72% of your plan allowance.
Illustrative responses — your account returns live production data.
Security & permissions
Query-only keys let your AI assistant read projects, strings, glossary terms, and usage — but never publish pages, edit live copy, or change billing. Keys use HTTPS and can be revoked instantly.
Issue one key per teammate, rotate when someone leaves, and keep write operations in the ConveyThis app or REST API. Questions? Centre d'aide.
Supported AI tools
Fonctionne avec tout client MCP prenant en charge le HTTP streamable et les jetons porteurs.
- Cursor
- Claude Code
- VS Code
- MCP Inspector
- Cline
- Windsurf
MCP vs Traditional Translation APIs
Mêmes données, deux interfaces : MCP pour les vérifications conversationnelles dans votre IDE ; API REST pour l'automatisation scriptée au moment du déploiement.
| Aspect | REST API | MCP Server AI tools |
|---|---|---|
| How you interact | Write HTTP requests, parse JSON, build scripts | Ask in plain language — your AI assistant handles the rest |
| Context for LLMs | Fetch and paste data into the chat yourself | Live translation context pulled automatically |
| Security model | Depends on key scopes you configure | Query-only keys — no publish or billing access |
| Developer experience | Best for CI pipelines and production automation | Best for ad-hoc checks and localization Q&A in the IDE |
| Learning curve | Requires API docs and code | Copy-paste setup in Cursor, Claude Code, or VS Code |
MCP complements visual editing and multilingual SEO — it doesn’t replace them.
Who it’s for
Enterprise use cases
Même configuration MCP — équipes différentes, questions différentes.
Engineering
Engineering & content ops
Catch untranslated routes before release and verify glossary coverage from the IDE.
“Any missing translations on /pricing?”
Agencies
Agency & multi-site portfolios
Comparez WordPress, Shopify, et les sites personnalisés à partir d'un seul compte ConveyThis.
“Which client site is least translated?”
Stay secure
Best practices
One key per person
Issue separate keys per teammate or environment. Revoke instantly when someone leaves.
Start narrow
Progress and glossary questions return faster than broad “scan everything” prompts.
Right tool for the job
API REST pour les pipelines de déploiement. MCP pour les vérifications de localisation ad hoc dans votre éditeur.
Feed the glossary
Les termes de marque verrouillés dans votre glossaire ConveyThis se retrouvent directement dans les réponses de l'IA.
Pourquoi ConveyThis
Built by a translation platform you already trust
ConveyThis powers multilingual sites on WordPress, Shopify, and enterprise properties worldwide — purpose-built for translation data, not a generic chatbot plugin.
MCP gives your AI assistant a safe window into glossary rules, project status, and production strings — on the same infrastructure behind millions of translated page views.
200+
languages supported
Query-only
keys by default
Production
translation infrastructure
FAQ
What is an MCP Server?
A backend service that speaks the Model Context Protocol — a standard way for AI assistants to discover tools and read structured data. The ConveyThis MCP Server exposes your projects, glossary, progress, and usage so Claude Code, Cursor, VS Code, and other clients can answer questions in plain language.
What is Model Context Protocol?
Model Context Protocol (MCP) is an open standard for connecting AI assistants to external data. Instead of copying exports into a chat window, your assistant calls the ConveyThis server over HTTP and gets live translation context.
Quel est le serveur MCP ConveyThis ?
Un serveur MCP en lecture seule conçu pour la gestion des traductions. Votre assistant IA peut lire les projets, la progression, les chaînes, les termes de la glossary, les règles et l'utilisation — en utilisant une clé API personnelle que vous créez sous MCP et agents IA.
How does Claude use MCP?
Claude Code registers the ConveyThis endpoint in its MCP configuration. When you ask about progress or glossary terms, Claude calls the server and composes an answer from live data. See our connect guide for copy-paste config.
Does Cursor support MCP?
Yes. Cursor supports Streamable HTTP MCP with bearer-token auth. Add the ConveyThis server to your settings, paste your key, and ask about translations directly in the IDE.
VS Code peut-il se connecter à un serveur MCP ?
Yes. VS Code extensions that implement MCP — including Cline — connect to https://mcp.conveythis.com/mcp with your bearer token. One endpoint works across Cursor, Claude Code, VS Code, and the MCP Inspector.
Can ChatGPT connect to MCP?
Any client that supports remote MCP connectors can use the same ConveyThis endpoint and bearer token. One-click OAuth for ChatGPT is on our roadmap.
Is the server read-only?
Oui. Les agents peuvent lire les données de traduction mais ne peuvent pas publier, modifier, supprimer ou changer la facturation. Révoquez une clé instantanément sous MCP et agents IA lorsque vous faites pivoter les identifiants.
Is translation data secure?
Keys are query-only, transmitted over HTTPS, and revocable at any time. Issue separate keys per developer for easier auditing.
En quoi le MCP est-il différent d'une API ?
Une API REST nécessite des points de terminaison, des paramètres et des scripts. MCP permet à votre assistant IA de transformer des questions en langage naturel en appels corrects — idéal pour les vérifications de localisation ad hoc sans écrire de code.
Comment obtenir une clé API pour le serveur MCP ?
Open MCP / AI Agents in ConveyThis, create a key (shown once), and paste it into your client as a bearer token. See the connect guide for per-client steps.
Quels clients d'IA peuvent se connecter à ConveyThis via MCP ?
Tout client avec MCP HTTP streamable et jetons porteurs — y compris Claude Code, Cursor, VS Code, MCP Inspector, Cline et Windsurf. Même point de terminaison et clé sur tous les outils.
Do I need programming knowledge?
No. Paste a config block, ask a question in English, and get an answer. Developers get tighter IDE integration; localization managers get plain-language progress and glossary lookups.
L'utilisation du serveur MCP entraîne-t-elle des frais supplémentaires ?
Non. L'accès MCP est inclus dans votre plan ConveyThis sans frais supplémentaires.
What is the MCP server endpoint URL?
https://mcp.conveythis.com/mcp — connectez-vous via Streamable HTTP avec votre clé comme en-tête Authorization: Bearer.