n8n-nodes-yc Package Description

Integration package for working with Yandex Cloud services in n8n.

Table of Contents

1. Yandex Cloud GPT Chat Model
2. Yandex Cloud Foundation Models
3. Yandex Object Storage
4. Yandex Cloud Functions
5. Yandex Cloud Containers
6. Yandex Cloud Compute
7. Yandex Cloud Data Streams
8. Yandex Cloud Message Queue
9. Yandex Cloud Postbox
10. Yandex Cloud SpeechKit
11. Yandex Cloud SpeechKit STT
12. Yandex Cloud Translate
13. Yandex Cloud Workflows
14. Yandex Cloud YDB
15. Yandex Cloud Search
16. Yandex Cloud Vision OCR
17. Yandex Cloud Logging

---

Yandex Cloud GPT Chat Model

Node for integration with Yandex Cloud Foundation Models API as a language model for LangChain. This node is an AI chain component and represents a Chat Model compatible with the LangChain framework. Designed for advanced use in AI agents and processing chains (AI Chain, AI Agent). The node has no inputs and outputs a language model object that can be connected to other AI components in n8n.

Parameter	Type	Description
Model	Resource Locator	YandexGPT model ID (e.g., yandexgpt/latest)
Maximum Tokens	Number	Maximum number of tokens in response (up to 8000)
Temperature	Number	Controls randomness of generation (0-1, default 0.6)
Timeout	Number	Request timeout in milliseconds (default 60000)
Max Retries	Number	Maximum number of retry attempts (default 2)

Authentication: Uses API key via yandexCloudGptApi credentials with folder ID passed in x-folder-id header. Supports base API URL configuration. The node is compatible with OpenAI API protocol, allowing use of standard LangChain tools. Ideal for creating chatbots, RAG systems, and AI agents using Russian language models.

---

Yandex Cloud Foundation Models

Node for working with Yandex Cloud Foundation Models API, providing direct access to YandexGPT generative language models via REST API. Unlike the Chat Model version, this node works as a regular transformation node with Main-type inputs and outputs, allowing requests to the API and receiving JSON responses in the context of standard workflows.

Parameter	Type	Description
Resource	Options	Resource type (Chat)
Operation	Options	Operation to perform

Resources and operations:

• Chat - work with model chat functionality

Authentication: Uses API key via yandexCloudFomoApi credentials. Supports base URL configuration and automatic HTTP error handling (ignoreHttpStatusErrors). Features model selection from dropdown list (similar to LMChatYandexGpt) and enhanced error messages with detailed API responses. The node is suitable for simple text generation scenarios that don't require complex LangChain integration but need direct work with Yandex Cloud API for creating dialogs, generating content, and natural language processing in automated workflows.

---

Yandex Object Storage

Node for comprehensive management of Yandex Cloud object storage, compatible with Amazon S3 API. Provides an exhaustive set of operations for working with both buckets (containers) and objects (files), using AWS SDK to ensure maximum compatibility and reliability.

Parameter	Type	Resources	Operations
Bucket	Resource	Bucket	Create, Delete, Get, List, Set ACL, Set Versioning
Object	Resource	Object	Upload, Download, Copy, Move, Delete, Get, List, Set ACL, Get Presigned URL

Bucket Operations:

• Create - create new bucket with ACL configuration
• Delete - delete empty bucket
• Get - get bucket information (location, metadata)
• List - enumerate all available buckets
• Set ACL - configure access rights (private, public-read, public-read-write, authenticated-read)
• Set Versioning - enable/disable object versioning

Object Operations:

• Upload - upload from binary data, text, or JSON with content-type, storage class (Standard/Cold/Ice) and metadata configuration. Returns objectUrl field containing the full object URL (e.g., https://storage.yandexcloud.net/my-bucket/file.txt)
• Download - download object to binary format
• Copy - copy between buckets preserving metadata. Returns objectUrl field with the destination object URL
• Move - move with automatic source deletion. Returns objectUrl field with the destination object URL
• Delete - delete object
• Get - get metadata (size, content-type, ETag, version)
• List - enumerate objects with prefix filtering and pagination (up to 1000 objects)
• Set ACL - configure access rights at object level
• Get Presigned URL - generate temporary signed URLs for secure access

Authentication: Uses static access keys (access key ID and secret access key) via yandexCloudStaticApi credentials. Supports resource locator for convenient bucket selection from list or manual name entry. The node is ideal for backup, user file storage, CDN integrations, and organizing data lakes in the cloud.

---

Yandex Cloud Functions

Node for invoking serverless functions in Yandex Cloud Functions via HTTP endpoints. Allows launching cloud functions with parameter passing and getting execution results, automatically managing authentication via IAM tokens.

Parameter	Type	Description
Folder ID	String	Folder ID for function search (default from credentials)
Function	Options	Select function from list or specify ID
HTTP Method	Options	GET or POST
Request Body	String (JSON)	Request body for POST method
Query Parameters	Collection	Collection of query parameters (name/value)
Headers	Collection	Additional HTTP headers

Execution process:

1. Automatic loading of available functions list from specified folder via SDK
2. Getting IAM token from service account credentials
3. Forming HTTP request to endpoint https://functions.yandexcloud.net/{functionId}
4. Adding Authorization header with Bearer token
5. Executing request and parsing response (automatic JSON detection)

Returned data:

• statusCode - HTTP response status
• headers - response headers
• body - response body (object or string)

Authentication: Uses service account JSON via yandexCloudAuthorizedApi credentials for automatic IAM token retrieval. Validates JSON request body before sending. The node is ideal for integrating business logic written in Python, Node.js, Go, and other languages into n8n workflows, enabling complex computations and data processing in serverless architecture.

---

Yandex Cloud Containers

Node for invoking serverless containers in Yandex Cloud Serverless Containers, providing the ability to run containerized applications via HTTP. Functionality is similar to Yandex Cloud Functions, but works with Docker containers, allowing use of any dependencies and environments.

Parameter	Type	Description
Folder ID	String	Folder ID for container search
Container	Options	Select container from list or specify ID
HTTP Method	Options	GET or POST
Request Body	String (JSON)	Request body for POST method
Query Parameters	Collection	URL request parameters
Headers	Collection	Custom HTTP headers

Workflow:

1. Loading list of available containers via SDK
2. Getting container URL (field url)
3. Generating IAM token for authentication
4. Executing HTTP request to container URL
5. Processing and returning response

Returned data:

• statusCode - HTTP response code
• headers - response headers
• body - response body (automatic JSON parsing)

Authentication: Service account JSON with automatic IAM token retrieval via yandexCloudAuthorizedApi. Supports JSON validation before sending. Differs from Functions by using full Docker images, providing more flexibility in runtime, library, and system dependency choices. Suitable for running ML models, complex applications with many dependencies, and microservices within n8n workflows.

---

Yandex Cloud Compute

Node for managing virtual machines in Yandex Compute Cloud, providing basic start and stop instance operations. Uses official Yandex Cloud SDK for interacting with Compute API.

Parameter	Type	Description
Resource	Options	Resource type (Instance)
Operation	Options	Start or Stop
Folder ID	String	Folder ID with virtual machines
Instance	Options	Select VM from list or specify ID

Operations:

• Start - start stopped virtual machine
• Stop - stop running virtual machine

Execution process:

1. Parsing service account JSON credentials
2. Creating SDK session with authentication
3. Getting instance list from specified folder
4. Executing start/stop operation via InstanceServiceClient
5. Returning operation information

Returned data:

• success - execution status
• operation - operation type (start/stop)
• instanceId - virtual machine ID
• operationId - operation ID in Yandex Cloud
• done - whether operation is completed
• metadata - operation metadata

Authentication: Service account JSON via yandexCloudAuthorizedApi with mandatory field validation (serviceAccountId, accessKeyId, privateKey). Node shows VM status when selecting from list (RUNNING, STOPPED, etc.). Ideal for infrastructure management automation: VM startup on schedule, stopping to save costs, integration with monitoring and alerting systems.

---

Yandex Cloud Data Streams

Node for working with stream data processing via Yandex Cloud Data Streams (YDS), compatible with Apache Kafka and Amazon Kinesis API. Provides high-performance data transfer between applications with guaranteed delivery and message ordering.

Parameter	Type	Resources	Operations
Record	Resource	Record	Put, Put Multiple
Stream	Resource	Stream	Describe, List

Record Operations:

• Put - send single record with data type selection (String/JSON), partition key specification and optional parameters (explicit hash key, sequence number)
• Put Multiple - batch record sending in two modes:
  • Define Records - manual record definition via UI
  • Use Input Data - automatic use of incoming items with field mapping

Stream Operations:

• Describe - get detailed stream information (status, retention period, shards, encryption)
• List - enumerate all available streams with limit

Send parameters:

• streamName - stream name (format: /ru-central1/{folder-id}/{database-id}/{stream-name})
• data - data to send (string or JSON)
• partitionKey - key for shard determination
• explicitHashKey - explicit hash specification for routing
• dataField - data field when using input data
• partitionKeyField - partition key field

Returned data:

• For Put: shardId, sequenceNumber, encryptionType
• For Put Multiple: successCount, failedCount, detailed information per record
• For Describe: full stream information, shards with hash ranges
• For List: list of stream names

Authentication: Static keys via yandexCloudStaticApi, uses Kinesis-compatible endpoint. Supports resource locator for convenient stream selection. The node is suitable for building real-time analytics, streaming ETL processing, log and metrics collection, microservice integration with guaranteed message delivery and horizontal scaling capability via sharding.

---

Yandex Cloud Message Queue

Node for sending messages to Yandex Cloud Message Queue (YMQ) queues, fully compatible with Amazon SQS API. Provides reliable asynchronous message passing between distributed system components with support for FIFO and standard queues.

Parameter	Type	Description
Resource	Options	Message
Operation	Options	Send
Queue	Resource Locator	Select queue from list or specify URL
Message Body	String	Message content (up to 256 KB)

Additional fields:

• Delay Seconds - message delivery delay (0-900 seconds)
• Message Deduplication ID - deduplication token (required for FIFO)
• Message Group ID - message grouping (required for FIFO)

Message Attributes:

Collection of message attributes with configuration:

• name - attribute name
• dataType - data type (String, Number, Binary)
• value - attribute value

Returned data:

• messageId - unique message ID
• md5OfMessageBody - MD5 hash of message body
• md5OfMessageAttributes - MD5 hash of attributes
• sequenceNumber - sequence number (for FIFO)
• success - send status
• queueUrl - queue URL

Authentication: Static keys via yandexCloudStaticApi, uses endpoint https://message-queue.api.cloud.yandex.net. Supports both standard queues (at-least-once delivery, best-effort ordering) and FIFO queues (exactly-once processing, strict ordering). The node is ideal for building event-driven architectures, splitting monoliths into microservices, buffering load between components, background task processing, and ensuring fault tolerance through asynchronous communication.

---

Yandex Cloud Postbox

Node for sending transactional emails via Yandex Cloud Postbox using Amazon SES v2 API. Provides reliable email delivery with HTML template and variable support, with guaranteed IP address reputation from Yandex Cloud.

Parameter	Type	Description
Resource	Options	Email
Operation	Options	Send
Email Type	Options	Simple or Template
From Email	String	Sender email (domain must be verified)
To Email	String	Recipient email (multiple comma-separated)

Simple Email (simple mode):

• Subject - email subject
• HTML Body - HTML version of email
• Text Body - text version for non-HTML clients

Template Email (template mode):

• Template Subject - subject with {{variable}} placeholders
• Template HTML - HTML template with placeholders
• Template Text - text version of template (optional)
• Template Data - JSON object with substitution data

Send process:

1. Parsing recipient list (comma separation)
2. Forming Content structure (Simple or Template)
3. Substituting variables in template (if used)
4. Sending via SES API with UTF-8 charset
5. Getting Message ID

Returned data:

• messageId - unique sent email ID
• success - send status
• from - sender address
• to - array of recipient addresses
• subject - email subject
• emailType - email type (simple/template)

Authentication: Static keys via yandexCloudStaticApi, uses endpoint https://postbox.cloud.yandex.net. Requires prior domain verification in Yandex Cloud. Supports multiple recipients and templating for personalization. The node is suitable for sending notifications, registration confirmations, password resets, reports, marketing campaigns, and any transactional emails with high deliverability and detailed analytics.

---

Yandex Cloud SpeechKit

Node for speech synthesis (Text-to-Speech) using Yandex SpeechKit API v3, providing high-quality audio generation from text. Supports multiple voices, emotional roles, and flexible synthesis parameter configuration.

Parameter	Type	Description
Resource	Options	Speech
Operation	Options	Synthesize
Text	String	Text for synthesis (multiline)
Voice	Options	Voice selection (Alena, Filipp, Jane, John, Masha, Omazh, Zahar)
Role	Options	Emotional coloring (Neutral, Good, Evil)

Available voices:

• Alena - female, Russian
• Filipp - male, Russian
• Jane - female, Russian/English
• John - male, English
• Masha - female, Russian
• Omazh - female, Russian
• Zahar - male, Russian

Audio formats:

Container (containerized files):

• WAV - uncompressed format
• MP3 - lossy compression
• OGG Opus - open codec

Raw PCM (raw audio stream):

• Sample rates: 8000, 16000, 22050, 48000 Hz
• Encoding: LINEAR16_PCM

Additional options:

• Speed - speech speed (0.1 - 3.0, default 1.0)
• Volume - volume (-145 to 1)
• Pitch Shift - voice pitch change (-1000 to 1000 Hz)

Synthesis process:

1. Parsing service account credentials
2. Creating SDK session
3. Connecting to TTS API (tts.api.cloud.yandex.net:443)
4. Streaming audio chunks generation
5. Combining chunks into final file
6. Preparing binary data with correct MIME type

Returned data:

• success - synthesis status
• text - source text
• voice - used voice
• role - emotional role
• audioFormat - audio format
• audioSize - file size in bytes
• Binary data - audio file with .wav/.mp3/.ogg/.raw extension

Authentication: Service account JSON via yandexCloudAuthorizedApi with automatic IAM token generation. Uses gRPC streaming for efficient audio transfer. The node is ideal for creating voice assistants, notification voiceovers, audiobook generation, accessibility solutions, IVR systems, and any applications requiring text-to-natural-speech conversion in Russian and English.

---

Yandex Cloud SpeechKit STT

Node for speech recognition (Speech-to-Text) using Yandex SpeechKit API v3, providing asynchronous audio transcription with high accuracy. Supports multiple languages and audio formats with automatic language detection capability.

Parameter	Type	Description
Operation	Options	Recognize Audio or Get Recognition Results
Audio URL	String	URL of audio file in Yandex Object Storage
Language Code	Options	Language for recognition (auto-detect, Russian, English, etc.)
Format Specification	Options	Choose to specify format using Audio Format codes or MIME Types
Audio Format	Options	Yandex Cloud audio format (LPCM, OGG Opus, MP3) - shown when Format Specification is "Audio Format"
MIME Type	Options	Standard MIME type (audio/wav, audio/ogg, audio/mpeg, etc.) - shown when Format Specification is "MIME Type"

Operations:

• Recognize Audio - start asynchronous audio transcription

  • Returns operation ID for polling
  • Supports audio files stored in Yandex Object Storage

• Get Recognition Results - retrieve transcription results with auto-polling

  • Automatically polls until completion
  • Handles "operation not ready" race condition
  • Configurable polling interval and max attempts
  • Optional partial results on timeout

Supported Languages:

• Automatic Detection
• Russian (ru-RU)
• English (en-US)
• German (de-DE)
• French (fr-FR)
• Spanish (es-ES)
• Italian (it-IT)
• Polish (pl-PL)
• Turkish (tr-TR)
• And many more (16 languages total)

Audio Formats:

The node supports two ways to specify audio format:

Via Audio Format (Yandex Cloud format codes):

• LPCM - Linear PCM with configurable sample rate (8000, 16000, 48000 Hz)
• OGG Opus - Compressed Ogg Opus format
• MP3 - Standard MP3 format

Via MIME Type (standard MIME types, automatically mapped to Yandex formats):

• audio/wav, audio/pcm, audio/x-pcm - Maps to LPCM
• audio/ogg, audio/opus - Maps to OGG Opus
• audio/mpeg, audio/mp3 - Maps to MP3

Recognition Options:

• Audio Channel Count - number of audio channels (1-8)
• Sample Rate - for LPCM format (8000, 16000, 48000 Hz)
• Profanity Filter - filter out profane language
• Literature Text - use literary text normalization

Polling Options (Get Results):

• Poll Interval - time between polling attempts (1-60 seconds, default 5)
• Max Attempts - maximum polling attempts (1-300, default 60)
• Return Partial Results - return incomplete results on timeout

Recognition Process:

1. Upload audio file to Yandex Object Storage
2. Start recognition with audio URL and parameters
3. Receive operation ID
4. Poll for results (auto-retry on "not ready" errors)
5. Get transcribed text with confidence scores

Returned data (Recognize):

• success - operation start status
• operationId - ID for polling results
• audioUrl - URL of audio file
• model - recognition model used (general)
• languageCode - selected language
• status - operation status (RUNNING)

Returned data (Get Results):

• success - completion status
• operationId - operation ID
• status - DONE or RUNNING
• text - full transcribed text
• channelTag - audio channel identifier
• finalResults - detailed results with alternatives
• attemptsUsed - number of polling attempts

Authentication: Service account JSON via yandexCloudAuthorizedApi with automatic IAM token generation. Uses gRPC streaming for efficient data transfer. Automatically handles race conditions when operation data is not immediately available. The node is ideal for creating voice-controlled interfaces, meeting transcription, call center analytics, subtitle generation, accessibility features, and any applications requiring accurate speech-to-text conversion in multiple languages with async processing support.

---

Yandex Cloud Translate

Node for text translation using Yandex Cloud Translate API, providing high-quality neural machine translation for 100+ languages. Supports automatic language detection, custom glossaries, HTML translation, and spell checking before translation.

Parameter	Type	Description
Resource	Options	Text or Language
Operation	Options	Translate, Detect Language, or List
Folder ID	String	Folder ID (default from credentials)

Resources and operations:

• Text - text translation and language detection
  • Translate - translate text to target language with optional glossary
  • Detect Language - automatically detect the language of text
• Language - supported languages management
  • List - get all supported languages

Translation parameters:

• texts - text to translate (multiline, supports multiple texts separated by newlines)
• sourceLanguageCode - source language (leave empty for auto-detection)
• targetLanguageCode - target language (required)
• format - text format (Plain Text or HTML)
• speller - enable spell checking before translation
• model - custom translation model ID (optional)
• glossary - custom glossary terms for precise translation

Glossary configuration:

Custom terminology dictionary for accurate translation of specific terms:

• sourceText - term in source language
• translatedText - term in target language
• exact - use exact word matching

Language Detection:

• text - text for language detection
• languageCodeHints - comma-separated language codes to prioritize (e.g., en,ru,es)

Supported Languages:

100+ languages including: Russian, English, German, French, Spanish, Italian, Portuguese, Polish, Turkish, Chinese, Japanese, Korean, Arabic, Hebrew, Hindi, and many more. Full list available via List Languages operation.

API Limits:

• Maximum 10,000 characters per request
• Multiple texts can be translated in single request
• Supports batch translation via newline-separated input

Returned data (Translate):

• success - translation status
• sourceLanguageCode - source language (auto-detected if not specified)
• targetLanguageCode - target language
• translations - array of translation results
  • text - translated text
  • detectedLanguageCode - detected source language

Returned data (Detect Language):

• success - detection status
• text - analyzed text
• languageCode - detected language code

Returned data (List Languages):

• Array of supported languages with:
  • code - language code (ISO 639-1)
  • name - language name

Use cases:

• Multi-language content localization
• Real-time chat translation
• Document translation workflows
• User-generated content translation
• Customer support in multiple languages
• API response localization
• Email and notification translation
• Website content translation

Authentication: Service account JSON via yandexCloudAuthorizedApi with automatic IAM token generation. Uses Yandex Cloud Translate API v2 with gRPC protocol for efficient communication. The node is ideal for building multilingual applications, automating content translation, creating translation pipelines, and integrating machine translation into business workflows with support for custom terminology and HTML content preservation.

---

Yandex Cloud Workflows

Node for starting workflow executions in Yandex Cloud Workflows, a serverless service for orchestrating cloud resources and microservices. Allows integrating n8n workflows with Yandex Workflows, creating hybrid automations.

Parameter	Type	Description
Resource	Options	Workflow
Operation	Options	Start Execution
Folder ID	String	Folder ID with workflows
Workflow	Options	Select workflow from list or specify ID
Input Data	String (JSON)	Input data for workflow in JSON format

Launch process:

1. Loading list of available workflows from specified folder
2. Validating input data (JSON correctness check)
3. Creating SDK client ExecutionServiceClient
4. Sending execution start request with specification of:
   • workflowId - workflow ID
   • input.inputJson - JSON string with input parameters
5. Getting execution ID

Returned data:

• executionId - unique execution ID
• workflowId - workflow ID
• success - launch status

Usage examples:

• Starting complex cloud function orchestration from n8n
• Data passing between n8n and Yandex Workflows
• Triggering long-running processes
• Integration with existing Workflows definitions

Authentication: Service account JSON via yandexCloudAuthorizedApi for creating SDK session. Automatically validates JSON before sending. Supports resource locator for convenient workflow selection with description display. The node is suitable for creating complex automations where n8n manages external integrations and APIs, while Yandex Workflows coordinates cloud resources (Functions, Containers, Compute), providing visual design of complex business processes with error handling, retry logic, and parallel task execution.

---

Yandex Cloud YDB

Node for working with Yandex Database (YDB), a distributed SQL database with automatic horizontal scaling and high availability. Provides ability to execute YQL (Yandex Query Language) queries with parameter binding, automatic type conversion, and support for multiple result sets.

Parameter	Type	Description
Resource	Options	Query
Operation	Options	Execute or Execute with Parameters
YQL Query	String	YQL query to execute
Return Mode	Options	All Result Sets, First Result Set, or First Row Only

Operations:

• Execute - run simple YQL query without parameters
• Execute with Parameters - run parameterized query with secure parameter binding

Query Parameters (for Execute with Parameters):

Collection of query parameters with configuration:

• name - parameter name (without $ prefix)
• value - parameter value (automatically converted to appropriate YDB type)

Return Modes:

• All Result Sets - returns all result sets from multi-statement queries
  • Returns: resultSets (array of result sets), resultSetCount, totalRows
• First Result Set - returns only the first result set as an array
  • Returns: rows (array of rows), rowCount
• First Row Only - returns only the first row of the first result set
  • Returns: single object or null

Type Conversion:

Automatic bidirectional conversion between JavaScript and YDB types:

JavaScript	YDB Type	Example
string	Text/String	"hello" → Text
number (int)	Int32	42 → Int32
number (float)	Double	3.14 → Double
bigint	Int64	9007199254740991n → Int64
boolean	Bool	true → Bool
Date	Datetime	new Date() → Datetime
Array	List	[1,2,3] → List
Object	Struct	{a:1} → Struct<a:Int32>
null	Null/Optional	null → Optional

Query Examples:

Simple Query:

SELECT id, name, email FROM users WHERE active = true LIMIT 10

Parameterized Query:

SELECT * FROM orders WHERE user_id = $userId AND order_date >= $startDate

Parameters: userId = 12345, startDate = "2025-01-01"

Batch Insert:

UPSERT INTO products SELECT * FROM AS_TABLE($rows)

Parameters: rows = array of product objects

Multiple Result Sets:

SELECT COUNT(*) as total_users FROM users;
SELECT COUNT(*) as total_orders FROM orders;
SELECT COUNT(*) as total_products FROM products;

Authentication:

The YDB node uses a dual credential approach for better security and flexibility:

1. Yandex Cloud Authorized API (Required)

   • Provides Service Account JSON for authentication
   • Shared across multiple Yandex Cloud services
   • Generates IAM tokens for secure access

2. Yandex Cloud YDB API (Required)

   • Provides YDB-specific connection parameters (Endpoint and Database)
   • Separates connection details from authentication
   • Allows easy switching between databases (dev/staging/prod)
   • Reusable across nodes with same database

This separation enables using one service account with multiple YDB databases while maintaining clear security boundaries.

Execution Process:

1. Parse service account JSON credentials
2. Generate IAM token via Yandex Cloud IAM service
3. Create YDB driver with authentication
4. Convert JavaScript parameters to YDB types using @ydbjs/value
5. Execute YQL query via @ydbjs/query
6. Convert YDB result sets back to JavaScript objects
7. Return data according to selected return mode
8. Close driver connection

Returned Data:

Depends on return mode:

• All Result Sets: {resultSets: [[...], [...]], resultSetCount: 2, totalRows: 15}
• First Result Set: {rows: [{id: 1, name: "Alice"}, ...], rowCount: 10}
• First Row Only: {id: 1, name: "Alice", email: "[email protected]"} or null

Use Cases:

• Real-time analytics queries
• Data aggregation and reporting
• User data management (CRUD operations)
• Complex joins and transformations
• ETL workflows with structured data
• Microservices data layer
• Event sourcing and audit logs
• Multi-tenant SaaS databases
• Transactional workloads with strong consistency

Best Practices:

• Always use parameterized queries for dynamic values (prevents SQL injection)
• Use AS_TABLE for batch operations (more efficient than individual inserts)
• Leverage YDB indexes for better query performance
• Monitor query statistics via return data
• Handle errors gracefully with Continue on Fail mode
• Consider using First Row Only for aggregates and counts
• Use YDB-specific features like UPSERT for idempotent operations

Usage: Requires both yandexCloudAuthorizedApi (service account authentication) and yandexCloudYdbApi (connection parameters). Uses Yandex Cloud IAM for token generation and @ydbjs SDK for database connectivity. The node is ideal for building data-driven applications, analytics dashboards, user management systems, and any scenarios requiring a distributed SQL database with horizontal scaling, strong consistency, and built-in replication in n8n workflows.

---

Yandex Cloud Search

Node for working with Yandex Cloud Search API, providing web search and AI-powered generative search capabilities. Uses official Yandex Cloud SDK to access search services with streaming support for generative responses and automatic XML parsing for web search results.

Parameter	Type	Description
Operation	Options	Web Search or Generative Search
Folder ID	String	Folder ID (default from credentials)

Operations:

Web Search

Perform traditional web search queries with extensive filtering and customization options.

Parameter	Type	Description
Query	String	Search query text (required)
Search Type	Options	RU, TR, COM, KK, BY, UZ (default: RU)

Additional Options:

• Family Mode - Content filtering (None, Moderate, Strict)
• Page Number - Results page (0-based, default: 0)
• Fix Typos - Auto-correct typos (Auto, Disabled)
• Sort By - Sort by relevance or time
• Sort Order - Ascending or descending (for time sorting)
• Group By - Flat or by domain
• Groups on Page - Number of groups per page (default: 10)
• Docs in Group - Documents per group (default: 1)
• Max Passages - Maximum passages in snippet (default: 2)
• Region - Search region ID (e.g., 213 for Moscow)
• Localization - Interface language (RU, UK, BE, KK, TR, EN)
• Response Format - XML or HTML (default: XML)
• Parse XML to JSON - Automatically parse XML response (default: true)

Returns:

• rawData - Raw XML/HTML response from search API
• parsedData - Parsed JSON structure (when parseXml is enabled)
• parseError - Error message if XML parsing fails

Generative Search

AI-powered search with natural language answers generated from search results.

Parameter	Type	Description
Messages	Collection	Conversation messages with role (User/Assistant) and content

Additional Options:

• Search Type - RU, TR, COM, KK, BY, UZ
• Site Restriction - Limit search to specific site (e.g., example.com)
• Host Restriction - Limit to specific host (e.g., <www.example.com>)
• URL Restriction - Limit to specific URL path
• Fix Misspell - Correct query misspells automatically
• Enable NRFM Docs - Include documents not on front page
• Get Partial Results - Stream partial results during processing

Returns:

• responses - Array of all streaming responses received
• finalResponse - Last response with complete answer
  • message - Generated answer text
  • sources - Array of source documents used (url, title, used flag)
  • searchQueries - Refined queries used for generation
  • fixedMisspellQuery - Corrected query if misspells were fixed
  • isAnswerRejected - Whether answer was rejected for ethical reasons
  • isBulletAnswer - Whether answer is in bullet-point format

Conversation Example:

{
  "messages": [
    { "role": "USER", "content": "What is n8n?" },
    { "role": "ASSISTANT", "content": "n8n is a workflow automation tool..." },
    { "role": "USER", "content": "How do I install it?" }
  ]
}

Execution Process:

1. Parse service account JSON credentials with validation
2. Create authenticated SDK session
3. Execute search operation via appropriate service client
4. For Web Search: parse XML response to JSON (if enabled)
5. For Generative Search: accumulate streaming responses
6. Return structured results with pairedItem references

Key Features:

• Streaming Support - Generative Search uses server-side streaming for progressive results
• XML Parsing - Automatic conversion of XML responses to JSON for easier processing
• Multi-Message Context - Support for conversational search with history
• Flexible Filtering - Extensive options for refining search results
• Source Attribution - Track which sources were used in generative answers
• Error Resilience - Graceful error handling with Continue on Fail support

Authentication: Uses service account JSON via yandexCloudAuthorizedApi credentials. Validates all required credential fields (serviceAccountId, accessKeyId, privateKey) and folder ID. Supports folder ID override at node level. The node is ideal for building search-powered applications, research tools, content discovery systems, AI assistants with web knowledge, and any scenarios requiring search integration with customizable relevance and filtering in n8n workflows.

---

Yandex Cloud Vision OCR

Node for text recognition in images using Yandex Cloud Vision OCR API, providing high-accuracy optical character recognition with support for multiple languages and formats. Supports JPEG, PNG, and PDF files with automatic MIME type detection and flexible output formatting.

Parameter	Type	Description
Resource	Options	Text Recognition
Operation	Options	Recognize
Binary Property	String	Name of the binary property containing image data (default: "data")
MIME Type	Options	Auto-detect, JPEG, PNG, or PDF
Languages	Multi-options	Languages to recognize (ISO 639-1 format)
Model	Options	OCR model selection (Page, Table, Markdown, Math Markdown)
Output Format	Options	Full Text Only, Structured Data, or Both

Supported Languages:

The OCR supports 50+ languages organized into two models:

Latin-Cyrillic Model:

• English, Russian, German, French, Spanish, Italian, Portuguese, Dutch, Polish
• Czech, Slovak, Bulgarian, Serbian, Bosnian, Croatian, Romanian, Hungarian
• Swedish, Norwegian, Danish, Finnish, Estonian, Latvian, Lithuanian
• Turkish, Azerbaijani, Kazakh, Uzbek, Tajik, Kyrgyz, Tatar, Bashkir, Yakut, Chuvash
• Indonesian, Maltese, Slovenian

Other Models (with Russian and English support):

• Arabic, Chinese, Japanese, Korean, Thai, Vietnamese
• Hebrew, Greek, Armenian, Georgian

All languages use ISO 639-1 format codes (e.g., 'en', 'ru', 'de')

OCR Models:

General text recognition models:

• Page - General purpose text recognition for documents and images (default)
• Page Column Sort - Multi-column text recognition
• Handwritten - Optimized for handwritten text recognition
• Table - Optimized for recognizing tabular data and structured layouts
• Markdown - Returns recognized text in markdown format
• Math Markdown - Specialized for mathematical formulas and equations in markdown

Template document recognition models:

• Passport - Template recognition for passport documents
• Driver License Front - Template recognition for driver license (front side)
• Driver License Back - Template recognition for driver license (back side)
• Vehicle Registration Front - Template recognition for vehicle registration certificate (front)
• Vehicle Registration Back - Template recognition for vehicle registration certificate (back)
• License Plates - Recognition of vehicle license plates

File Format Support:

• JPEG - Standard JPEG images
• PNG - PNG images with transparency support
• PDF - Single-page PDF documents
• Maximum file size: 10MB
• Maximum image dimensions: 20M pixels (width × height)

Output Formats:

1. Full Text Only - Simple string containing all recognized text

   • Returns: { fullText: "recognized text..." }

2. Structured Data - Detailed structure with coordinates and metadata

   • Returns: { structured: [{ page, width, height, blocks, entities, tables, markdown, pictures }] }
   • Includes bounding boxes for blocks, lines, and words
   • Provides table cell recognition with row/column indices
   • Detects entities and pictures with confidence scores

3. Both - Combined output with both full text and structured data

   • Returns: { fullText: "...", structured: [...] }

Recognition Process:

1. Get image data from specified binary property
2. Auto-detect or use provided MIME type
3. Validate file size (max 10MB)
4. Validate model requirements (e.g., license-plates requires explicit language)
5. Send recognition request with language preferences and model selection
6. Stream response chunks from OCR API
7. Combine results for multi-page PDFs
8. Format output according to selected format

Returned Data (Full Text):

• fullText - Complete recognized text from all pages

Returned Data (Structured):

• structured - Array of page annotations
  • page - Page number (1-based)
  • width / height - Image dimensions in pixels
  • blocks - Text blocks with bounding boxes, lines, words, languages
  • entities - Named entities detected in text
  • tables - Recognized tables with cell data (row/column indices, spans, text)
  • markdown - Text in markdown format (if model supports)
  • pictures - Detected picture locations with confidence scores
  • rotate - Image rotation angle (ANGLE_0, ANGLE_90, ANGLE_180, ANGLE_270)

Structured Data Details:

Block structure:

• boundingBox - Polygon vertices defining block area
• lines - Array of text lines with words and orientation
• languages - Detected languages in block
• layoutType - Block type (TEXT, HEADER, FOOTER, TITLE, LIST, etc.)

Table structure:

• rowCount / columnCount - Table dimensions
• cells - Array of table cells with:
  • rowIndex / columnIndex - Cell position
  • rowSpan / columnSpan - Cell spanning
  • text - Cell content

Use Cases:

• Document digitization and archiving
• Invoice and receipt processing
• ID card and passport data extraction
• Table data extraction from images
• Scanned document text extraction
• Multi-language document processing
• Form processing automation
• OCR for accessibility features
• Content moderation with text detection
• License plate recognition (with structured coordinates)

MIME Type Detection:

The node automatically detects image format by analyzing magic bytes:

• JPEG: FF D8 FF header
• PNG: 89 50 4E 47 header
• PDF: 25 50 44 46 header
• Falls back to JPEG if detection fails

Error Handling:

• Validates binary data presence and size
• Checks file size limit (10MB)
• Validates model-specific requirements (e.g., license-plates needs explicit language)
• Validates service account credentials
• Supports Continue on Fail mode for batch processing
• Returns detailed error messages with context

Multi-page PDF Support:

For PDF files with multiple pages, the node:

• Processes each page separately
• Combines full text with double newline separators
• Returns structured data array with page numbers
• Preserves page-specific metadata and coordinates

Authentication: Uses service account JSON via yandexCloudAuthorizedApi credentials with automatic IAM token generation. Uses gRPC streaming for efficient data transfer. Connects to ocr.api.cloud.yandex.net:443 for recognition requests. The node is ideal for building document processing workflows, data extraction pipelines, form automation, invoice processing systems, and any applications requiring accurate text recognition from images with support for complex layouts, tables, and mathematical notation in n8n workflows.

---

Yandex Cloud Logging

Node for centralized log management with Yandex Cloud Logging service, providing write and read operations for log entries. Supports structured logging with JSON payloads, time-based filtering, and resource-based organization with automatic pagination and batch operations.

Parameter	Type	Resources	Operations
Log Entry	Resource	Log Entry	Write, Read

Log Entry Operations:

• Write - send log entries to log group with:

  • Multiple entries in single request
  • Log levels (TRACE, DEBUG, INFO, WARN, ERROR, FATAL)
  • Automatic or custom timestamps
  • JSON payloads for structured data
  • Stream names for log organization
  • Resource metadata (type and ID)
  • Default values for level, payload, and stream

• Read - retrieve log entries with filtering:

  • Time range filtering (since/until)
  • Log level filtering (multi-select)
  • Resource type and ID filtering
  • Stream name filtering
  • Advanced filter expressions
  • Automatic pagination support
  • Return all entries or limit results

Write Parameters:

• logGroupId - target log group (resource locator with dropdown)
• entries - array of log entries with:
  • message - log message text (required)
  • level - log level (TRACE/DEBUG/INFO/WARN/ERROR/FATAL)
  • timestamp - ISO 8601 format or auto-generated
  • jsonPayload - structured data as JSON object
  • streamName - stream for log organization
• resourceType / resourceId - optional resource metadata
• defaults - default level, payload, and stream for all entries

Read Parameters:

• logGroupId - source log group (resource locator)
• returnAll - fetch all pages or limit results
• limit - maximum entries to return (1-1000)
• filters - optional filtering:
  • since / until - time range (ISO 8601)
  • levels - log levels to include
  • resourceTypes - comma-separated resource types
  • resourceIds - comma-separated resource IDs
  • streamNames - comma-separated stream names
  • filter - advanced filter expression

Returned Data (Write):

• success - operation status
• entriesWritten - number of entries written
• logGroupId - target log group ID
• errors - map of any entry-level errors

Returned Data (Read):

Each log entry returned as separate item with:

• uid - unique entry ID
• message - log message
• level - log level
• timestamp - entry timestamp
• resource - resource metadata (type, id)
• jsonPayload - structured data
• streamName - stream name
• ingestedAt - ingestion timestamp
• savedAt - save timestamp

Use Cases:

• Centralized application logging
• Error monitoring and alerting
• Audit trail and compliance logging
• Distributed system log aggregation
• Real-time log analysis pipelines
• Performance monitoring and metrics
• Security event logging
• Troubleshooting and debugging

Features:

• Batch Writing - send multiple entries in single request
• Structured Logging - JSON payloads for rich data
• Time-Based Filtering - precise time range queries
• Resource Organization - group logs by resource type/ID
• Stream Support - organize logs by application component
• Automatic Pagination - handle large result sets
• Resource Locator - easy log group selection from dropdown

Authentication: Uses service account JSON via yandexCloudAuthorizedApi credentials with automatic IAM token generation. Requires logging.writer role for writing entries and logging.reader role for reading entries. The node is ideal for building observability systems, monitoring dashboards, log analysis workflows, alerting pipelines, and any applications requiring centralized structured logging with powerful filtering and querying capabilities in n8n workflows.

---

Authentication Types

The package uses four types of credentials:

yandexCloudYdbApi

• Endpoint (YDB endpoint URL)
• Database (YDB database path)
• Used for Yandex Cloud YDB connection parameters (requires yandexCloudAuthorizedApi for authentication)

yandexCloudGptApi

• API key for Foundation Models
• Folder ID
• Endpoint URL (optional)

yandexCloudStaticApi

• Access Key ID
• Secret Access Key
• Used for S3-compatible services (Object Storage, Data Streams, Message Queue, Postbox)

yandexCloudAuthorizedApi

• Service Account JSON
• Folder ID
• Used for services requiring IAM tokens (Functions, Containers, Compute, SpeechKit, Workflows, YDB)

---

Common Features

All nodes in the package support:

• ✅ Continue on Fail - continue execution on errors
• ✅ Paired Items - maintain relationship between input and output items
• ✅ Resource Locators - convenient resource selection from lists or manual entry
• ✅ Expressions - use n8n expressions in all parameters
• ✅ Proxy Support - work via HTTP/HTTPS proxy (where applicable)