Flixu文
Feature Spotlight

Translate from inside your AI assistant. No dashboard required.

Flixu's Agent Skills let you trigger localization directly from Claude Code, Cursor, or any MCP-compatible AI assistant. Type what you want in plain English. Flixu handles the translation, applies your brand voice and glossary, and returns the result — without switching tools.

What are Flixu AI Agents?

Flixu AI Agents expose translation operations as callable skills inside AI assistants that support MCP (Model Context Protocol). Instead of exporting files, logging into a dashboard, and waiting for a translation to run, a developer can describe the task to their AI assistant and Flixu executes it in the background — with your Translation Memory, Glossary, and Brand Voice already applied.

What the workflow looks like without this.

For most developers using AI coding assistants today, localization is still a manual step that happens outside the AI workflow. The code gets written in English, the new strings get extracted manually, uploaded to a localization tool, translated, downloaded, and merged back. That handoff breaks the flow.

Even teams with automated CI/CD localization through the GitHub App still interact with a dashboard when they want to translate a specific string, check a translation, or run a quick localization test before a commit.

The Agent Skills integration removes that interruption. Localization becomes a capability your AI assistant already has — one you invoke with a sentence, not a tab switch.

How Flixu Agent Skills work.

Flixu provides an MCP (Model Context Protocol) server that exposes translation operations as tools. Any AI assistant that supports MCP — including Claude Code and Cursor — can connect to Flixu and call those tools directly. The connection is set up once. After that, it works like this:

1

Describe the task in natural language

You type your instruction directly in your AI assistant:

"Translate the new checkout flow strings into French using our enterprise brand voice"

or

"Localize these three UI labels into German, Japanese, and Spanish — use the SaaS product glossary"

The AI assistant passes the request to Flixu through the MCP connection.

2

Flixu runs the full translation pipeline

The request hits Flixu's translation pipeline with your workspace context already loaded — the Translation Memory, Glossary, and Brand Voice associated with your account or client profile. The pre-translation analysis runs, the language model generates output with all constraints applied, and the LQA scores the result. This is the same pipeline that runs on a full project document. The agent interface doesn't bypass it.

3

Result returns to your workflow

The translated strings return to your AI assistant's context. You can review them, ask for adjustments, pass them to a file, or commit them — all without leaving the environment where you were already working.

What you can ask Flixu to do from inside your AI assistant.

The scope of what Agent Skills handle is the same as what Flixu handles in a project — the difference is the interface:

  • — Translate individual strings or short copy with a specified client profile
  • — Translate larger content blocks with brand voice and glossary enforcement
  • — Apply a specific language target and formality level to a translation request
  • — Check a translation against glossary constraints
  • — Run a quick translation test before pushing strings to the full CI/CD pipeline

What it doesn't do yet: automate full multi-file project workflows entirely from the AI assistant interface. For large-scale automated localization, the GitHub App handles that path. The Agent Skills integration is the interface for developer-initiated, in-context translation tasks during the build process.

→ For fully automated pipeline localization: GitHub Integration

What this looks like in practice.

A developer adds a new payment confirmation flow to a checkout page. The flow has six UI strings. They're working in Claude Code and have already been coding for two hours. Rather than switching to Flixu, exporting the strings, and running a project, they type:

Developer prompt

Translate these checkout confirmation strings into German and French.
Use the e-commerce brand voice profile. Apply the product glossary.

strings:

  • - "Your order is confirmed."
  • - "You'll receive a confirmation email shortly."
  • - "Track your order"
  • - "Continue shopping"
  • - "Order #{{order_id}}"
  • - "Estimated delivery: {{date}}"

Flixu's MCP server receives the request, loads the specified brand voice and glossary, runs the pre-translation analysis, and returns:

Flixu response

German:

  • - "Ihre Bestellung ist bestätigt."
  • - "Sie erhalten in Kürze eine Bestätigungs-E-Mail."
  • - "Bestellung verfolgen"
  • - "Weiter einkaufen"
  • - "Bestellung #{{order_id}}"
  • - "Voraussichtliche Lieferung: {{date}}"

French:

  • - "Votre commande est confirmée."
  • - "Vous recevrez bientôt un e-mail de confirmation."
  • - "Suivre votre commande"
  • - "Continuer vos achats"
  • - "Commande #{{order_id}}"
  • - "Livraison estimée : {{date}}"

Variables are preserved. Glossary terms are applied. Brand voice is reflected in the output register. The developer pastes the strings directly into their localization files. The whole interaction took under two minutes and never left the coding environment.

When the agent integration changes the workflow.

Developer-led teams using AI coding assistants daily

For teams where the developer writes code and manages localization without a dedicated localization manager, the context switch to a localization dashboard is a real friction point. AI Agent integration reduces localization to a task the developer can handle from their existing AI assistant context — the same way they'd ask for a code review or a function refactor.

Teams that move from dashboard-based ad-hoc translation to agent-based in-context translation find the adoption rate for localization increases — developers use the capability more because the barrier is lower.

→ Developer workflow overview: Flixu for Developers

Micro-SaaS founders managing multiple product areas solo

A solo founder working on a product in multiple languages uses every tool at once. Context switching to a localization dashboard for each new string adds overhead that compounds across a day. Being able to invoke translation directly from the AI assistant they're already using — without a separate login, without file management — reduces that overhead to a conversational instruction.

→ Indie founder and micro-SaaS workflows: Developer API

Technical documentation teams using AI-assisted writing

Documentation teams using AI assistants to draft content can invoke Flixu to translate completed documentation sections directly from the writing workflow. The Brand Voice Manager ensures documentation tone is consistent across languages — the same careful technical voice in English translates to the same careful technical voice in German or Japanese.

Where this is today.

The MCP integration is actively being developed and expanded. Current Agent Skills support in-context translation for strings and short-form content with workspace context applied. Multi-file project management and full workflow automation from within an AI assistant is the direction we're building toward — it's not yet the full scope of what's available.

If you're building workflows around this integration or want to test current capabilities, the Developer API documentation has the most current information on what's available and what's coming. The GitHub App covers the automated CI/CD localization path for larger-scale pipeline automation today.

Frequently Asked Questions

What is MCP and why does it matter for localization?

+

MCP (Model Context Protocol) is a standard developed by Anthropic that lets AI assistants connect to external tools and services through a defined interface. When an AI assistant supports MCP, it can call external tools — like Flixu's translation pipeline — as part of its natural operation. For developers, this means translation becomes a capability their AI assistant already has rather than a separate workflow they have to switch to.

Which AI assistants support Flixu's Agent Skills?

+

Any AI assistant that supports MCP can connect to Flixu's MCP server. This includes Claude Code and Cursor. As MCP adoption grows across AI coding tools, the range of compatible assistants expands. Check the Developer API documentation for the current list of tested integrations.

Does the agent apply my Brand Voice and Glossary automatically?

+

Yes — when you specify a client profile or workspace context in the request, the Agent Skills integration uses the same Brand Voice and Glossary configuration as a full project translation. The translation pipeline doesn't bypass those constraints because the request came through an AI assistant rather than the dashboard.

How is this different from the GitHub App?

+

The GitHub App handles automated localization triggered by repository events — a developer pushes new strings, Flixu detects them, translates, and commits automatically. The Agent Skills integration is for developer-initiated, in-context tasks during the build process — translating strings on demand from inside an AI assistant, testing translations before a commit, or handling ad-hoc localization tasks that don't need to be part of an automated pipeline.

Do translations run through Agent Skills count toward my credit balance?

+

Yes. Translations invoked through Agent Skills use the same credit system as translations run through the dashboard or GitHub App. Credits = words translated, regardless of which interface triggers the request.

Can I use Agent Skills for non-string content — documentation, emails, product copy?

+

Yes. Agent Skills handle string-level translation and short-to-medium content. For long-form structured documents, the full document translation pipeline (via the dashboard or Developer API) handles format preservation and segment-level LQA scoring. Agent Skills work well for in-context translation tasks where you need an answer in the next minute, not a full project workflow.

Connect Flixu to your AI assistant.

Set up the MCP integration once and invoke translation directly from Claude Code, Cursor, or any MCP-compatible assistant. Your Brand Voice and Glossary apply automatically.

Developer API and integration docs →

Related Features