Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 2 additions & 5 deletions ar/basic/admin-panel/ai-setting.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -48,12 +48,9 @@ description: "هيئ محادثة الذكاء الاصطناعي وحقوله

في **LLM API**، أضف موفّر نماذج واختبر نماذجه. إذا كنت تستخدم مفتاحك الخاص (BYOK)، فاختر **متوافق مع OpenAI** أو **Anthropic** بوصفه نوع الموفّر.

افتح **إعدادات النموذج** لأحد النماذج عندما تحتاج إلى مراجعة القيم التالية أو تجاوزها:
تُحدَّد نافذة السياق وأقصى إخراج في الاستجابة الواحدة تلقائيًا من **كتالوج النماذج**، ولا يلزم إدخالها يدويًا. ولمراجعة القيم المحددة لنموذج ما، افتح **كتالوج النماذج**، ثم اختر **النطاق** في **فحص النموذج** وحدد النموذج؛ عندئذٍ تعرض **تفاصيل المعاملات** السياق وأقصى إخراج ومصدر كل قيمة.

| الإعداد | التأثير |
|------|------|
| **السياق** | اختياري. يحدد الحد الأقصى لرموز الإدخال المستخدمة في ضغط السياق. |
| **الحد الأقصى للمخرجات** | اختياري. يحدد الحد الأقصى للرموز التي يمكن للنموذج توليدها في استجابة واحدة. |
وإذا لم يتضمن الكتالوج مدخلًا مطابقًا، كما في نموذج صدر حديثًا أو نموذج خاص بك، فإن النموذج يعمل بالقيم المحددة في **لا تطابق** ضمن **قواعد المطابقة**: نافذة سياق 200000 وأقصى إخراج 32000 افتراضيًا، ويظهر **قاعدة المطابقة** مصدرًا لكلتيهما في **تفاصيل المعاملات**. وإذا كانت هذه القيم لا تناسب النموذج، فعدّلها في **قواعد المطابقة** ثم اضغط **حفظ**.

### تهيئة النماذج الموصى بها

Expand Down
2 changes: 1 addition & 1 deletion ar/basic/ai/ai-chat.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ description: تفاعل مع بياناتك باللغة الطبيعية لتح
## عناصر التحكم في الإدخال

- **الإدخال الصوتي**: انقر على زر الميكروفون في مربع الإدخال واسمح بالوصول إلى الميكروفون لبدء التسجيل. أثناء التسجيل، اختر **إنهاء الإدخال الصوتي** أو **تجاهل الإدخال الصوتي**. بعد اختيار الإنهاء، ينسخ Teable التسجيل إلى نص في مربع الإدخال.
- **النموذج**: استخدم قائمة النماذج لاختيار مستوى النموذج للمحادثة. تعرض القائمة **Ultra** و**Smart** و**Standard** و**Lite** من الأعلى قدرة إلى الأدنى، مع بيان النموذج القائم خلف كل مستوى. ويمكن للمسؤول تبديل ذلك النموذج في أي وقت أو إتاحة بعض المستويات فقط، من دون أن تحتاج إلى إعادة الاختيار. استخدم مستوى أدنى للاستعلامات البسيطة أو التنظيف أو إعادة الصياغة، ومستوى أعلى للتخطيط المعقد والتحليل عبر الجداول وإنشاء التطبيقات.
- **النموذج**: استخدم قائمة النماذج لاختيار مستوى النموذج للمحادثة. تعرض القائمة **Ultra** و**Smart** و**Standard** و**Lite** من الأعلى قدرة إلى الأدنى، مع بيان النموذج القائم خلف كل مستوى. ويمكن للمسؤول تبديل ذلك النموذج في أي وقت أو إتاحة بعض المستويات فقط، من دون أن تحتاج إلى إعادة الاختيار. استخدم مستوى أدنى للاستعلامات البسيطة أو التنظيف أو إعادة الصياغة، ومستوى أعلى للتخطيط المعقد والتحليل عبر الجداول وإنشاء التطبيقات. وفي الإصدار السحابي، تعرض القائمة إلى جانب كل مستوى معامل قدرة الحوسبة مقارنةً بالمستوى الافتراضي (مثل **2.5×**)، ولا يظهر هذا المعامل عندما يستهلك المستوى قدرًا مماثلًا للمستوى الافتراضي أو لا يستهلك قدرة حوسبة المنصة. وعند تبديل المستوى أثناء المحادثة يُطلب تأكيد واحد، وتبقى المحادثة كما هي، ويتولى النموذج الجديد المعالجة ابتداءً من رسالتك التالية.
- **مستوى الذكاء**: اختر مستوى **الذكاء** بصورة مستقلة في قائمة النماذج. يتحكم هذا المستوى في عمق التفكير؛ وتنتج المستويات الأعلى استدلالًا أشمل.
- **أسراري**: افتح **+** ← **المزيد** ← **أسراري** لحفظ مفاتيح API أو غيرها من بيانات الاعتماد في **الإعدادات** ← **التكاملات**. هذه القيم ملك لك، ويمكن لمحادثة الذكاء الاصطناعي قراءتها أثناء التشغيل بوصفها متغيرات بيئة، كما يمكنك منحها لتطبيقات وأتمتة بعينها. راجع [بيانات الاعتماد وعمليات التكامل](/ar/basic/credential).
- **بطاقات التفويض**: عندما يحتاج الذكاء الاصطناعي إلى حساب خارجي أو إلى سرّ، يثبّت بطاقة في المحادثة وينتظر تصرّفك قبل أن تُكمل الجولة. فإذا كان Cuppy نفسه يحتاج الحساب لإنجاز المهمة، اضغط **اتصال** لتنفيذ تصريح OAuth؛ وإذا كان تطبيق أو أتمتة يحتاج بيانات الاعتماد أثناء التشغيل، اختر **ربط حسابي** أو امنح بيانات اعتماد تملكها بالفعل. وإذا نقرت على **تخطي** فلن تُقدَّم بيانات الاعتماد، ويواصل الذكاء الاصطناعي الأجزاء التي يمكنه إكمالها من دونها. راجع [بيانات الاعتماد وعمليات التكامل](/ar/basic/credential).
Expand Down
7 changes: 2 additions & 5 deletions de/basic/admin-panel/ai-setting.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -48,12 +48,9 @@ Klicken Sie auf **Test ausführen**, um die AI Chat-Laufzeit zu überprüfen.

Fügen Sie unter **LLM API** einen Modellanbieter hinzu und testen Sie dessen Modelle. Wenn Sie Ihren eigenen Schlüssel (BYOK) verwenden, wählen Sie **OpenAI Compatible** oder **Anthropic** als Anbietertyp.

Öffnen Sie für ein Modell die **Modelleinstellungen**, wenn Sie diese Werte prüfen oder überschreiben müssen:
Kontextfenster und maximale Ausgabe pro Antwort ermittelt der **Modellkatalog** automatisch; eine manuelle Eingabe entfällt. Um die für ein Modell ermittelten Werte zu prüfen, öffnen Sie **Modellkatalog**, wählen unter **Modell prüfen** einen **Bereich** und anschließend das Modell; **Parameterdetails** listet dann Kontext, max. Ausgabe und die Herkunft jedes Werts auf.

| Einstellung | Wirkung |
|------|------|
| **Kontext** | Optional. Legt die maximale Anzahl an Eingabe-Tokens für die Kontextkomprimierung fest. |
| **Maximale Ausgabe** | Optional. Legt die maximale Anzahl an Tokens fest, die das Modell in einer Antwort generieren kann. |
Enthält der Katalog keinen passenden Eintrag (etwa bei einem neu veröffentlichten oder selbst betriebenen Modell), gelten für das Modell die Werte aus **Nichts getroffen** in **Zuordnungsregeln**: standardmäßig ein Kontextfenster von 200000 und eine maximale Ausgabe von 32000, wobei **Parameterdetails** für beide **Zuordnungsregel** als Herkunft ausweist. Passen Sie die Werte in **Zuordnungsregeln** an und klicken Sie auf **Speichern**, wenn sie nicht zum Modell passen.

### Empfohlene Modelle konfigurieren

Expand Down
2 changes: 1 addition & 1 deletion de/basic/ai/ai-chat.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ Wenn Sie einen langen Block mit Klartext einfügen, wandelt Teable ihn in einen
## Eingabesteuerung

- **Spracheingabe**: Klicken Sie im Eingabefeld auf die Mikrofontaste und erlauben Sie den Mikrofonzugriff, um die Aufnahme zu starten. Wählen Sie während der Aufnahme **Spracheingabe beenden** oder **Spracheingabe verwerfen**. Nach dem Beenden transkribiert Teable die Aufnahme in das Eingabefeld.
- **Modell**: Verwenden Sie das Modellmenü, um die Modellstufe für die Unterhaltung auszuwählen. Das Menü führt **Ultra**, **Smart**, **Standard** und **Lite** von der leistungsstärksten Stufe abwärts auf und zeigt zu jeder das dahinterliegende Modell. Ihre Administration kann dieses Modell jederzeit austauschen oder nur einen Teil der Stufen anbieten, ohne dass Sie erneut wählen müssen. Nutzen Sie eine niedrigere Stufe für einfache Abfragen, Bereinigung oder Überarbeitungen und eine höhere Stufe für komplexe Planung, tabellenübergreifende Analysen und App-Erstellung.
- **Modell**: Verwenden Sie das Modellmenü, um die Modellstufe für die Unterhaltung auszuwählen. Das Menü führt **Ultra**, **Smart**, **Standard** und **Lite** von der leistungsstärksten Stufe abwärts auf und zeigt zu jeder das dahinterliegende Modell. Ihre Administration kann dieses Modell jederzeit austauschen oder nur einen Teil der Stufen anbieten, ohne dass Sie erneut wählen müssen. Nutzen Sie eine niedrigere Stufe für einfache Abfragen, Bereinigung oder Überarbeitungen und eine höhere Stufe für komplexe Planung, tabellenübergreifende Analysen und App-Erstellung. In der Cloud weist das Menü zu jeder Stufe außerdem den Faktor der Rechenleistung gegenüber der Standardstufe aus (zum Beispiel **2.5×**); bei einer Stufe, die etwa so viel wie die Standardstufe oder gar keine Rechenleistung der Plattform kostet, entfällt die Angabe. Ein Stufenwechsel mitten in einer Unterhaltung wird einmal bestätigt; die Unterhaltung bleibt erhalten, und das neue Modell übernimmt ab Ihrer nächsten Nachricht.
- **Intelligenz**: Wählen Sie die Stufe **Intelligenz** separat im Modellmenü. Sie steuert die Denktiefe; höhere Stufen liefern gründlichere Überlegungen.
- **Meine Secrets**: Öffnen Sie **+** → **Mehr** → **Meine Secrets**, um API-Schlüssel oder andere Anmeldedaten unter **Einstellungen** → **Integrationen** zu hinterlegen. Die Werte gehören Ihnen, der KI-Chat kann sie zur Laufzeit als Umgebungsvariablen lesen, und Sie können sie einzelnen Apps und Automatisierungen freigeben. Siehe [Zugangsdaten und Integrationen](/de/basic/credential).
- **Autorisierungskarten**: Wenn die KI ein Drittanbieterkonto oder ein Secret benötigt, heftet sie eine Karte an die Unterhaltung und wartet, bevor der Durchlauf weitergeht. Braucht Cuppy das Konto für die Aufgabe selbst, starten Sie mit **Verbinden** den OAuth-Vorgang; braucht eine App oder Automatisierung die Zugangsdaten zur Laufzeit, wählen Sie **Mein Konto verbinden** oder geben vorhandene Zugangsdaten frei. Mit **Überspringen** stellen Sie sie nicht bereit; die KI fährt mit den Teilen fort, die sie auch ohne erledigen kann. Siehe [Zugangsdaten und Integrationen](/de/basic/credential).
Expand Down
7 changes: 2 additions & 5 deletions en/basic/admin-panel/ai-setting.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -48,12 +48,9 @@ Click **Run test** to verify the AI Chat runtime.

In **LLM API**, add a model provider and test its models. If you use your own key (BYOK), choose **OpenAI Compatible** or **Anthropic** as the provider type.

Open **Model Settings** for a model when you need to review or override these values:
A model's context window and maximum output per response are resolved automatically from the **Model catalog**; there is nothing to fill in. To check the values resolved for a model, open **Model catalog**, choose a **Scope** under **Inspect model** and select the model; **Parameter details** then lists the context, max output, and the source of each value.

| Setting | Effect |
|------|------|
| **Context** | Optional. Sets the maximum input tokens used for context compaction. |
| **Max output** | Optional. Sets the maximum tokens the model can generate in one response. |
When the catalog holds no matching entry (a newly released or self-built model, for example), the model runs on the values set under **Nothing matched** in **Matching rules**: a context window of 200000 and a max output of 32000 by default, with **Matching rule** shown as the source of both in **Parameter details**. If those values do not fit the model, adjust them in **Matching rules** and click **Save**.

### Configure Recommended Models

Expand Down
2 changes: 1 addition & 1 deletion en/basic/ai/ai-chat.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ When you paste a long block of plain text, Teable converts it to a Markdown atta
## Input Controls

- **Voice input**: Click the microphone button in the input box and allow microphone access to start recording. While recording, choose **Finish voice input** or **Discard voice input**. After you choose finish, Teable transcribes the recording into the input box.
- **Model**: Use the model menu to choose the model tier for the conversation. The menu lists **Ultra**, **Smart**, **Standard**, and **Lite** from the most capable down, each showing the model behind it; your admin can swap that model at any time, or offer only some of the tiers, and you never have to pick again. Use a lower tier for simple queries, cleanup, or rewriting, and a higher tier for complex planning, cross-table analysis, and app building.
- **Model**: Use the model menu to choose the model tier for the conversation. The menu lists **Ultra**, **Smart**, **Standard**, and **Lite** from the most capable down, each showing the model behind it; your admin can swap that model at any time, or offer only some of the tiers, and you never have to pick again. Use a lower tier for simple queries, cleanup, or rewriting, and a higher tier for complex planning, cross-table analysis, and app building. In Cloud, the menu also marks each tier with its credit multiplier against the default tier (for example **2.5×**); a tier that costs about as much as the default, or no platform credit at all, shows no multiplier. Switching tiers during a conversation asks you to confirm once; the conversation is kept, and the new model takes over from your next message.
- **Intelligence**: Choose the **Intelligence** level separately in the model menu. It controls thinking depth; higher levels produce more thorough reasoning.
- **My secrets**: Open **+** → **More** → **My secrets** to store API keys or other credentials in **Settings** → **Integrations**. The values are yours, AI Chat can read them as environment variables at runtime, and you can grant them to individual apps and automations. See [Credentials and Integrations](/en/basic/credential).
- **Authorization cards**: When AI needs a third-party account or a secret, it pins a card to the conversation and waits for you before the turn goes on. When Cuppy needs the account to do the task itself, click **Connect** to run an OAuth flow; when an app or automation needs the credential at run time, choose **Connect my account** or grant a credential you already have. Click **Skip** to withhold it; AI continues with the parts it can complete without it. See [Credentials and Integrations](/en/basic/credential).
Expand Down
7 changes: 2 additions & 5 deletions es/basic/admin-panel/ai-setting.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -48,12 +48,9 @@ Haz clic en **Ejecutar prueba** para verificar el entorno de ejecución del Chat

En **API de LLM**, añade un proveedor de modelos y prueba sus modelos. Si usas tu propia clave (BYOK), elige **Compatible con OpenAI** o **Anthropic** como tipo de proveedor.

Abre **Ajustes del modelo** para un modelo cuando necesites revisar o sustituir estos valores:
La ventana de contexto y la salida máxima por respuesta se obtienen automáticamente del **Catálogo de modelos**; no hay que rellenar nada. Para comprobar los valores obtenidos para un modelo, abre **Catálogo de modelos**, elige un **Ámbito** en **Inspeccionar modelo** y selecciona el modelo; **Detalle de parámetros** mostrará el contexto, la salida máxima y el origen de cada valor.

| Ajuste | Efecto |
|------|------|
| **Contexto** | Opcional. Establece el número máximo de tokens de entrada utilizados para compactar el contexto. |
| **Salida máxima** | Opcional. Establece el número máximo de tokens que el modelo puede generar en una respuesta. |
Cuando el catálogo no tiene una entrada coincidente (por ejemplo, un modelo recién publicado o propio), el modelo funciona con los valores definidos en **Sin coincidencias**, dentro de **Reglas de coincidencia**: de forma predeterminada, una ventana de contexto de 200000 y una salida máxima de 32000, y **Detalle de parámetros** indica **Regla de coincidencia** como origen de ambos. Si esos valores no se ajustan al modelo, modifícalos en **Reglas de coincidencia** y pulsa **Guardar**.

### Configurar los modelos recomendados

Expand Down
2 changes: 1 addition & 1 deletion es/basic/ai/ai-chat.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ Cuando pegas un bloque largo de texto sin formato, Teable lo convierte en un arc
## Controles de entrada

- **Entrada de voz**: Haz clic en el botón del micrófono del cuadro de entrada y permite el acceso al micrófono para comenzar a grabar. Durante la grabación, elige **Finalizar entrada de voz** o **Descartar entrada de voz**. Al finalizar, Teable transcribe la grabación en el cuadro de entrada.
- **Modelo**: Usa el menú de modelos para elegir el nivel de modelo de la conversación. El menú lista **Ultra**, **Smart**, **Standard** y **Lite**, del más capaz al menos capaz, e indica el modelo que hay detrás de cada uno; tu administrador puede cambiar ese modelo en cualquier momento u ofrecer solo algunos niveles, y tú no tienes que volver a elegir. Usa un nivel más bajo para consultas sencillas, limpieza o reescritura, y uno más alto para planificación compleja, análisis entre Tablas y creación de aplicaciones.
- **Modelo**: Usa el menú de modelos para elegir el nivel de modelo de la conversación. El menú lista **Ultra**, **Smart**, **Standard** y **Lite**, del más capaz al menos capaz, e indica el modelo que hay detrás de cada uno; tu administrador puede cambiar ese modelo en cualquier momento u ofrecer solo algunos niveles, y tú no tienes que volver a elegir. Usa un nivel más bajo para consultas sencillas, limpieza o reescritura, y uno más alto para planificación compleja, análisis entre Tablas y creación de aplicaciones. En Cloud, el menú indica además en cada nivel el multiplicador de potencia respecto al nivel predeterminado (por ejemplo, **2.5×**); no aparece cuando el nivel consume aproximadamente lo mismo que el predeterminado o no consume potencia de la plataforma. Cambiar de nivel durante una conversación pide una confirmación; la conversación se mantiene y el nuevo modelo se hace cargo a partir de tu siguiente mensaje.
- **Inteligencia**: Elige el nivel de **Inteligencia** por separado en el menú de modelos. Este controla la profundidad del razonamiento; los niveles superiores producen razonamientos más exhaustivos.
- **Mis secretos**: Abre **+** → **Más** → **Mis secretos** para guardar claves de API u otras credenciales en **Ajustes** → **Integraciones**. Esos valores son tuyos, el Chat de IA puede leerlos como variables de entorno durante la ejecución y puedes concederlos a aplicaciones y automatizaciones concretas. Consulta [Credenciales e integraciones](/es/basic/credential).
- **Tarjetas de autorización**: Cuando la IA necesita una cuenta externa o un secreto, fija una tarjeta en la conversación y espera antes de seguir con el turno. Si es Cuppy quien necesita la cuenta para hacer la tarea, pulsa **Conectar** para iniciar un flujo de OAuth; si la credencial la necesita una aplicación o automatización al ejecutarse, elige **Conectar mi cuenta** o concede una credencial que ya tengas. Con **Omitir** no la facilitas; la IA continúa con las partes que puede completar sin ella. Consulta [Credenciales e integraciones](/es/basic/credential).
Expand Down
Loading
Loading