Spoken Arabic for AI agents: dialects, sentiment, Arabizi, names, Tamazight Tifinagh.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
API key from your plan. Optional: without it you get a daily free trial.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Sentiment and opinion analysis for customer reviews, messages and social media posts written in spoken Arabic dialects (Moroccan Darija, Egyptian, Gulf, Levantine and others), in Arabic script, Arabizi or mixed with French/English. Returns overall sentiment, a score from -1 to 1, the aspects mentioned (price, delivery, quality, service...) with their sentiment, the detected dialect and a one-line English summary. Built for review monitoring, social listening and customer support triage.
Review, comment or message (max 4000 characters).
Dialect of the text, if known. Default: auto-detect.
This tool isn't marked read-only, so it may change something on the server or in an account you connected.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http darija-mcp https://darija-mcp.lamtoueqali.workers.dev/mcp
Headers
X-API-Key (optional, not in the config above): API key from your plan. Optional: without it you get a daily free trial.Sentiment and opinion analysis for customer reviews, messages and social media posts written in spoken Arabic dialects (Moroccan Darija, Egyptian, Gulf, Levantine and others), in Arabic script, Arabizi or mixed with French/English. Returns overall sentiment, a score from -1 to 1, the aspects mentioned (price, delivery, quality, service...) with their sentiment, the detected dialect and a one-line English summary. Built for review monitoring, social listening and customer support triage.
| Argument | Type | Description |
|---|---|---|
| text* | string | Review, comment or message (max 4000 characters). |
| dialect | string | Dialect of the text, if known. Default: auto-detect. |
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "Review, comment or message (max 4000 characters)."
},
"dialect": {
"enum": [
"darija",
"algerian",
"tunisian",
"libyan",
"egyptian",
"sudanese",
"levantine",
"iraqi",
"gulf",
"yemeni"
],
"type": "string",
"description": "Dialect of the text, if known. Default: auto-detect."
}
}
}Given a person or family name of Arabic origin (in Arabic script or Latin letters), returns its Arabic spelling and the common Latin spelling variants used across countries (e.g. Mohammed / Muhammad / Mohamed / Mohamad), plus particle variants (Al-, El-, bin, ben, Abd, Abdel...). Use it to search records, deduplicate CRMs or broaden lookups. It is a search aid, not an identity or compliance decision.
| Argument | Type | Description |
|---|---|---|
| name* | string | Full or partial name (max 200 characters). |
| max_variants | integer | Max Latin variants per name part. Default: 10. |
{
"type": "object",
"required": [
"name"
],
"properties": {
"name": {
"type": "string",
"description": "Full or partial name (max 200 characters)."
},
"max_variants": {
"type": "integer",
"maximum": 30,
"minimum": 3,
"description": "Max Latin variants per name part. Default: 10."
}
}
}Converts Arabic dialect text written in Latin letters with digits (Arabizi / Franco-Arabic, e.g. "wach nta mzyan? 3afak" or "ezayak 3amel eh") into Arabic script, without translating it. Useful to clean social media, WhatsApp or customer-support data before analysis.
| Argument | Type | Description |
|---|---|---|
| text* | string | Arabizi text (max 4000 characters). |
| dialect | string | Dialect of the text, if known. Default: darija. |
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "Arabizi text (max 4000 characters)."
},
"dialect": {
"enum": [
"darija",
"algerian",
"tunisian",
"libyan",
"egyptian",
"sudanese",
"levantine",
"iraqi",
"gulf",
"yemeni"
],
"type": "string",
"description": "Dialect of the text, if known. Default: darija."
}
}
}FREE. Guesses which spoken Arabic dialect a text is written in (Moroccan, Algerian, Tunisian, Libyan, Egyptian, Sudanese, Levantine, Iraqi, Gulf, Yemeni), in Arabic script or Arabizi. Use it to pick the right "dialect" for the translation tools.
| Argument | Type | Description |
|---|---|---|
| text* | string | Text to analyse (max 4000 characters). |
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "Text to analyse (max 4000 characters)."
}
}
}FREE. Detects whether a text is Moroccan Darija (Arabic script or Arabizi/Latin) and returns a confidence score, the script used and the Darija markers found.
| Argument | Type | Description |
|---|---|---|
| text* | string | Text to analyse (max 4000 characters). |
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "Text to analyse (max 4000 characters)."
}
}
}FREE and instant. Transliterates Tamazight (Amazigh/Berber) between Latin script and Tifinagh (Neo-Tifinagh, IRCAM standard), e.g. "tamaziɣt" or "tamazight" ↔ ⵜⴰⵎⴰⵣⵉⵖⵜ. Direction is detected automatically. Accepts IRCAM spelling (ḍ ṣ ṭ ẓ ṛ ḥ ɛ ɣ) and common web spellings (gh, kh, ch, 3, 7). Letter-level transliteration, not translation.
| Argument | Type | Description |
|---|---|---|
| text* | string | Text in Latin or Tifinagh script (max 4000 characters). |
| direction | string | Default: auto. |
| emphatic_uppercase | boolean | Treat uppercase D, S, T, Z, R, H as emphatic consonants (ḍ, ṣ, ṭ, ẓ, ṛ, ḥ). Default: false. |
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "Text in Latin or Tifinagh script (max 4000 characters)."
},
"direction": {
"enum": [
"auto",
"to_tifinagh",
"to_latin"
],
"type": "string",
"description": "Default: auto."
},
"emphatic_uppercase": {
"type": "boolean",
"description": "Treat uppercase D, S, T, Z, R, H as emphatic consonants (ḍ, ṣ, ṭ, ẓ, ṛ, ḥ). Default: false."
}
}
}Translates spoken Arabic dialect text (Moroccan Darija by default, or Egyptian, Levantine, Gulf, etc.) in Arabic script, Arabizi or mixed with French/English, into English, French, Spanish, German, Italian, Turkish or Modern Standard Arabic. Handles slang and code-switching better than generic translators.
| Argument | Type | Description |
|---|---|---|
| text* | string | Dialect text (max 4000 characters). |
| target | string | Target language. Default: en. |
| dialect | string | Arabic dialect. Default: darija (Moroccan). Darija is native-reviewed; other dialects are in beta. |
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "Dialect text (max 4000 characters)."
},
"target": {
"enum": [
"en",
"fr",
"es",
"de",
"it",
"tr",
"msa"
],
"type": "string",
"description": "Target language. Default: en."
},
"dialect": {
"enum": [
"darija",
"algerian",
"tunisian",
"libyan",
"egyptian",
"sudanese",
"levantine",
"iraqi",
"gulf",
"yemeni"
],
"type": "string",
"description": "Arabic dialect. Default: darija (Moroccan). Darija is native-reviewed; other dialects are in beta."
}
}
}Translates English, French, Spanish, German, Italian, Turkish or Standard Arabic text into a natural spoken Arabic dialect, as natives really write it: Moroccan Darija by default, or Egyptian, Levantine, Gulf, Iraqi, Algerian, Tunisian and more. Arabic script, Arabizi (Latin) or both; casual, polite or business tone. Ideal for customer messages, marketing and chatbots in Arab markets.
| Argument | Type | Description |
|---|---|---|
| text* | string | Text to translate (max 4000 characters). |
| script | string | "latin" = Arabizi as used on WhatsApp and social media. Default: arabic. |
| dialect | string | Arabic dialect. Default: darija (Moroccan). Darija is native-reviewed; other dialects are in beta. |
| register | string | Tone. Default: polite. |
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "Text to translate (max 4000 characters)."
},
"script": {
"enum": [
"arabic",
"latin",
"both"
],
"type": "string",
"description": "\"latin\" = Arabizi as used on WhatsApp and social media. Default: arabic."
},
"dialect": {
"enum": [
"darija",
"algerian",
"tunisian",
"libyan",
"egyptian",
"sudanese",
"levantine",
"iraqi",
"gulf",
"yemeni"
],
"type": "string",
"description": "Arabic dialect. Default: darija (Moroccan). Darija is native-reviewed; other dialects are in beta."
},
"register": {
"enum": [
"casual",
"polite",
"business"
],
"type": "string",
"description": "Tone. Default: polite."
}
}
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 22, 2026.
No readOnlyHint / destructiveHint annotations, so clients can't tell a read from a write without asking.
Fix: Add annotations to each tool, e.g. { readOnlyHint: true } for tools that only read.