Pinecone

Get detailed information about a specific collection. Args: - collectionName: The name of the collection Returns the collection details including name, size, status, dimension, and vector count.

Get detailed information about a restore job. Args: - restoreId: The restore job ID Returns restore job details including status and progress.

Delete a collection. Args: - collectionName: The name of the collection to delete WARNING: This operation is irreversible.

List backups in the project. Args: - indexName: Optional - filter backups for a specific index Returns an array of backup objects with status, metadata, and size information.

Create a backup of an index. Args: - indexName: The name of the index to backup - name: Optional name for the backup - description: Optional description Returns the backup details.

Get detailed information about a specific backup. Args: - backupId: The backup ID Returns backup details including status, size, and metadata.

Delete a backup. Args: - backupId: The backup ID to delete WARNING: This operation is irreversible.

Create a new index from a backup. Args: - name: Name for the new index - backupId: The backup ID to restore from - deletionProtection: Optional deletion protection setting - tags: Optional custom tags Returns the new index configuration.

List all restore jobs in the project. Returns an array of restore job objects with status and progress information.

List all collections in the project. Returns an array of collection objects with name, size, status, dimension, and vector count.

Create a new collection from an existing index. Args: - name: Collection name (1-45 chars, lowercase alphanumeric or hyphens) - source: Name of the source index to create collection from The collection will be a static snapshot of the source index.

Start a bulk import operation from cloud storage. Args: - indexName: The name of the index to import into - uri: The URI of the data to import (e.g., s3://bucket/path/) - integrationId: Optional integration ID for cloud storage access - errorMode: How to handle errors ('abort' or 'continue') Returns the import job details.

List all import operations for an index. Args: - indexName: The name of the index Returns an array of import job objects with status and progress information.

Get detailed information about an import operation. Args: - indexName: The name of the index - importId: The import job ID Returns import job details including status, progress, and any errors.

Cancel a running import operation. Args: - indexName: The name of the index - importId: The import job ID to cancel Note: Vectors already imported before cancellation will remain in the index.

List all Pinecone indexes in the project. Returns an array of index objects with name, dimension, metric, host, spec, and status.

Create a new Pinecone index. Args: - name: Index name (1-45 chars, lowercase alphanumeric or hyphens) - dimension: Vector dimension (1-20000) - metric: Distance metric (cosine, euclidean, dotproduct). Default: cosine - cloud: Cloud provider (aws, gcp, azure) for serverless - region: Cloud region for serverless - environment: Pod environment (for pod-based indexes) - podType: Pod type like p1.x1 (for pod-based indexes) - pods: Number of pods (for pod-based indexes) - replicas: Number of replicas (for pod-based indexes) - deletionProtection: Enable deletion protection (enabled/disabled) Returns the created index configuration.

Get detailed information about a specific index. Args: - indexName: The name of the index Returns the index configuration including name, dimension, metric, host, spec, and status.

Delete a Pinecone index. Args: - indexName: The name of the index to delete WARNING: This operation is irreversible. All data in the index will be permanently deleted.

Update the configuration of an existing index. Args: - indexName: The name of the index to configure - replicas: Number of replicas (pod-based indexes only) - podType: Pod type (pod-based indexes only) - deletionProtection: Enable or disable deletion protection - tags: Custom tags to set Returns the updated index configuration.

Get statistics about an index. Args: - indexName: The name of the index - filter: Optional metadata filter to get stats for matching vectors only Returns dimension, total vector count, index fullness, and per-namespace statistics.

Generate vector embeddings for text inputs. Args: - model: The embedding model to use (e.g., 'multilingual-e5-large', 'llama-text-embed-v2') - inputs: Array of text strings to embed - inputType: Optional - 'query' or 'passage' for asymmetric models - truncate: How to handle long inputs - 'END' (truncate) or 'NONE' (error) Returns embeddings (arrays of numbers) for each input.

Rerank documents by relevance to a query. Args: - model: The reranking model to use (e.g., 'bge-reranker-v2-m3') - query: The query to rank documents against - documents: Array of document objects to rerank - topN: Number of top results to return (default: all) - returnDocuments: Include document content in response (default: true) - rankFields: Fields to consider for ranking (default: ['text']) Returns documents sorted by relevance score.

List available inference models. Returns information about available embedding and reranking models.

Get detailed information about a specific model. Args: - modelName: The name of the model Returns model details including supported parameters and configuration.

List all namespaces in an index. Args: - indexName: The name of the index Returns an array of namespaces with their vector counts.

Get information about a specific namespace. Args: - indexName: The name of the index - namespaceName: The name of the namespace Returns the namespace details including vector count.

Delete a namespace and all its vectors. Args: - indexName: The name of the index - namespaceName: The name of the namespace to delete WARNING: This will delete all vectors in the namespace.

Upsert vectors into an index. Args: - indexName: The name of the index - vectors: Array of vectors, each with id, values, and optional metadata - namespace: Optional namespace to upsert into Each vector should have: - id: Unique identifier (string) - values: Array of numbers (must match index dimension) - metadata: Optional key-value pairs Returns the count of upserted vectors.

Query an index for similar vectors. Args: - indexName: The name of the index - vector: Query vector (array of numbers) - topK: Number of results to return (1-10000) - namespace: Optional namespace to query - filter: Optional metadata filter - includeValues: Include vector values in response - includeMetadata: Include metadata in response - id: Alternative to vector - query by existing vector ID Returns matched vectors with scores.

Fetch vectors by their IDs. Args: - indexName: The name of the index - ids: Array of vector IDs to fetch - namespace: Optional namespace Returns the requested vectors with their values and metadata.

Update a vector's values or metadata. Args: - indexName: The name of the index - id: The vector ID to update - values: New vector values (optional) - setMetadata: Metadata to set or update (optional) - namespace: Optional namespace At least one of values or setMetadata must be provided.

Delete vectors from an index. Args: - indexName: The name of the index - ids: Array of vector IDs to delete (max 1000) - deleteAll: Delete all vectors in the namespace - filter: Delete vectors matching this metadata filter - namespace: Optional namespace Provide one of: ids, deleteAll, or filter.

List vector IDs in an index (serverless indexes only). Args: - indexName: The name of the index - namespace: Optional namespace - prefix: Optional ID prefix to filter by - limit: Maximum number of IDs to return (default 100) - paginationToken: Token for pagination Returns a list of vector IDs and pagination info.