Give the agent access to a spell checker by connecting it to an MCP server that exposes a narrowly scoped tool—such as check_spelling—and have that server call a checker such as LanguageTool. The agent can then request checks and receive structured findings, while your application keeps control over whether suggested edits are applied.
How the integration fits together
The setup has three parts:
- MCP client: The client attached to the AI agent discovers tools offered by MCP servers and makes them available to the model.
- Spell-check MCP server: This small wrapper declares a tool, validates its inputs, calls a checking service, and returns concise findings.
- Checking backend: A service such as LanguageTool checks the submitted text and returns issues, offsets, messages, and replacement candidates.
MCP provides the tool-discovery and invocation mechanism; it does not itself check spelling or prescribe a particular approval screen. The protocol describes servers exposing tools that language models can invoke. MCP Tools specification
As an Amazon Associate I earn from qualifying purchases.
This is an integration pattern based on MCP’s tool interface and LanguageTool’s HTTP API, not a claim that a particular MCP spell-check server or client has been tested.
Design a focused spell-check tool
Declare a tool named check_spelling with a clear description and a small input schema. A useful starting point is:
#1 Best Overall
- 【OCR Scan Translator】The reading pen supports text scanning in 55 languages. Translation pen with OCR recognition technology, the pen scanner can quickly scan words or sentences and read them aloud to you after scanning, which is applicable to books, e-books, newspapers, digital screens, labels, wood, etc. Important text can be transferred to the computer for editing via USB cable.
- 【Photo Translation and Smart Recording】The translation scanner pen is equipped with high-definition camera and a large touch screen. Just point your camera at any text and the Scan Reading Pen will automatically translate it. The scanning pen can also be used as a practical audio recorder to record and save all important interviews, meetings and conversations.
- 【Two Way Real-Time Voice Translation】The text to speech device supports online two-way real-time translation in 112 languages, response time is less than 0.3 seconds (faster than human translation). The reading pen scanner has an accurate detection rate of up to 98%, help you overcome cross-language barriers such as checking into hotels and visiting attractions when traveling abroad.
- 【Collins Dictionary】The scanning translation pen is equipped with authoritative Chinese English dictionaries from FLTRP and Collins Dictionary, supporting scanning of different fonts, making it your lightweight "dictionary" choice. The reader pen has a smaller appearance and is easy to carry to meet any mobile needs. Multipurpose translation pen is applicable to travel, study abroad and business travel.
- 【Widely Used】The pen dictionary supports 12 interface languages, with eye protecting UI. The reader pen even has a reverse scanning direction setting, which takes care of left-handed people very much. Reading pen is equipped with Bluetooth module, which can be connected to headphones or speakers. The text to speech device for dyslexia can be widely used in study, work, shopping and tourism.
text: the passage to check.language: a language or regional language code, such asen-USorde-DE.preferredVariants(optional): preferred regional variants when the backend’s automatic detection cannot distinguish them.
LanguageTool’s API accepts text and language settings. For variant languages, its documentation warns that spelling checks may not work with only a generic code such as en or de; use the intended variant when it matters. LanguageTool HTTP API reference
Keep the schema narrow instead of allowing the model to pass arbitrary backend parameters. Validate input size and language values in the server, and avoid forwarding unrelated options the agent does not need.
Rank #2
- 【Text to Voice】The scanning translator can scan 3,000 characters per minute, scan and translate the entire line of text within one second, and output the original text and translation by voice. The accuracy rate is as high as 98%, convenient and fast! Ideal for business work, student studies, and those with dyslexia. It is a good helper for learning foreign languages. It also supports offline use.
- 【150 Languages Voice Translator Pen】The voice translator supports online scan translation in 75 languages and real-time voice translation in 150 languages. Support multi-national accents, adjustable voice output speed. It is the best choice for you to take notes, record meetings, travel abroad, take exams, and give gifts.
- 【Two-way voice translation】This translation pen supports scanning and editing anytime, anywhere! Translations are instantly played through the built-in speaker and displayed on the pen, e.g. from Spanish to English or from English to Spanish.
- 【Offline Translation】Even when there is no network, the scanning translation pen also supports offline scanning and translation. The powerful Chinese-English electronic dictionary function is the best choice for you to learn English. 820mAh high-capacity battery supports up to 8 hours of continuous work and 7 days of standby time!
- 【Easy to Use】This instant language translation device features a 2.3-inch high-definition IPS screen and minimalist design. The simple operating system makes it easy for everyone to use it. Using the AI engine, combined with the proprietary neural network translation technology, it is not only fast, but also has a very high translation accuracy rate of over 98%.
Return findings, not invisible rewrites
Map the checker’s response into compact, structured findings. For each issue, retain the reported offset and length, the explanation, and any replacement candidates. Those fields let the client highlight the relevant span and present options without silently replacing the user’s text.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →If the input contains markup, preserve the relationship between the submitted text and the original document. LanguageTool supports structured text data that separates text from markup; that can help keep returned offsets meaningful when checking formatted content. LanguageTool HTTP API reference
Rank #3
- Improve English Proficiency: Help improve kids' English proficiency with this translation scanner. It enables preschool English enlightenment and independent reading, making it easy for children to learn English.
- Efficient Translation: The translation scanner supports full sentence and line breaking translation, ensuring efficient scanning and learning, and providing areading experience.
- Convenient and Practical: This smart scanning pen is designed specifically for students to scan and learn from Chinese and English books, making it convenient and practical for studying.
- Multiple Functions: With the ability to scan English, Chinese, ancient poetry, words, phrases, and long sentences, this smart scanning pen offers multiple functions. It features easy button and touch operation, making it a simple and practical tool.
- Long-lasting Battery: Built with a 650mAh lithium battery, this smart scanner pen offers long-lasting usage and can be easily charged using the Type C charging interface. It also comes with a data cable, allowing you to use it anytime, anywhere.
Treat results as suggestions. A misspelling may be a name, a deliberate spelling choice, or a valid term the checker does not recognize. Show uncertain corrections for review, and ask before making consequential edits. Tool output and annotations should be treated as untrusted unless the server is trusted; the MCP specification defines the tool mechanism, not a universal user-approval workflow. MCP Tools specification
Choose a backend: hosted API or local server
LanguageTool documents both a public HTTP API and an embedded HTTP server. The right choice depends on whether automated requests are permitted for the service and plan you select, what data may leave your environment, the checking capabilities you need, and who will operate the service.
| Consideration | Hosted public API | Self-hosted server |
|---|---|---|
| Automated use | LanguageTool says not to send automated requests to its public endpoint; it recommends setting up an instance or obtaining an Enterprise account for automation. Public API guidance | Runs as a server you set up and operate. Embedded HTTP Server |
| Data handling | Text is sent to the hosted service. Check the applicable service terms and privacy information; LanguageTool says users are responsible for informing their own users about text handling. Public API guidance | Can keep checking within your environment, depending on how you deploy and configure it. |
| Checking capabilities | LanguageTool’s setup documentation distinguishes cloud AI-based rules from the basic self-hosted server, which does not include those rules. Embedded HTTP Server | The basic self-hosted server does not include the cloud’s AI-based rules. Embedded HTTP Server |
| Operations | Service operation is handled by the provider, subject to the selected service’s limits and availability. | You are responsible for running, configuring, and updating the local service. Embedded HTTP Server |
Public endpoint limits and suitability
LanguageTool documents the public endpoint as https://api.languagetool.org/v2/check and specifies POST requests. Its documentation explicitly says, “Do not send automated requests.” The endpoint is therefore not the appropriate backend for an agent that will call it automatically; use a permitted service arrangement or a server you operate. LanguageTool Public HTTP Proofreading API
Recommended Free Tools
As stated in LanguageTool’s documentation accessed in 2026, the public service limits are 20 requests per IP per minute, 75 KB of text per IP per minute, 20 KB per request, and suggestions for up to 30 misspelled words. These are published service limits, not availability guarantees; the documentation says limits may change and that public-service versions and behavior may change without warning. LanguageTool Public HTTP Proofreading API
Best Value
- AI Reading Companion for Growing Readers – Help children stay engaged with every book. Scan difficult words, sentences, or ideas to get reading-level explanations with Quick AI. Tap Simplify for easier understanding, reduce frustration, and build the confidence to read independently.
- Instant Reading Support with Quick AI – Scan a word, phrase, or sentence to receive clear, age-appropriate explanations in seconds. Quick AI helps children understand difficult vocabulary, long sentences, and complex ideas without leaving the page or interrupting their reading.
- Explore Beyond the Page with AI Reading Buddy – Every scan automatically syncs to AI Reading Buddy for deeper learning. Discover the big ideas, get fun facts and insights beyond the scanned text with Tell Me More, and reinforce comprehension with AI-generated reading quizzes.
- More Than a Dictionary Pen – Includes offline tools like Text-to-Speech, translation, and a built-in dictionary for uninterrupted reading anywhere. When connected, unlock Quick AI, AI Reading Buddy, plus enhanced dictionaries and translation for deeper learning support.
- Designed to Build Independent Readers – Created for growing readers, ESL learners, and students who need extra reading support. WorldPenScan encourages curiosity, strengthens comprehension, and helps children develop the confidence to solve reading challenges on their own.
Before connecting a backend, check its permitted use, request-size and throughput limits, availability terms, language coverage, and data-handling terms. The public endpoint’s published limits should not be assumed to apply to a self-hosted deployment or another plan.
Quick Recap
Practical request flow
- Register the MCP server with the agent’s client. The client must support connecting to MCP servers and exposing their tools to the model.
- Have the client discover the tool. The server advertises
check_spelling, its description, and input schema through MCP. - Validate the request. Check that text is present, within the backend’s allowed size, and paired with an appropriate language setting.
- Call the selected checker. For LanguageTool’s documented public API, requests use POST, but its public endpoint must not be used for automated requests. Select an authorized backend for the agent’s automated calls. Public API guidance
- Return structured issues. Preserve offsets, lengths, messages, and replacement candidates so the application can present findings in context.
- Let the user decide about edits. Present suggestions and apply a replacement only through the application’s intended editing and confirmation flow.
Common integration mistakes
- Using a generic language code for a regional spelling choice: Set a regional variant such as
en-USorde-DEwhere appropriate. - Calling the public endpoint automatically: LanguageTool explicitly disallows automated requests to that public endpoint; use an authorized service or your own instance.
- Returning a rewritten document instead of findings: Keep issue locations and candidate corrections so users can inspect what would change.
- Losing offsets around formatting: Preserve markup-aware input handling when checking formatted documents.
- Assuming local and cloud checks are identical: LanguageTool says its basic self-hosted server lacks AI-based rules available in the cloud.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
