> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pinecone.io/llms.txt
> Use this file to discover all available pages before exploring further.

# 2025 releases

> Pinecone release notes — 2025 releases:

## December 2025

<Update label="2025-12-29" tags={["Assistant"]}>
  ### Increased metadata limit for assistants

  The metadata field limit for assistants has been increased from 1KB to 16KB.

  Assistant metadata is a JSON object that you can use to store custom organizational data, tags, and attributes for your assistants. You can specify metadata when creating an assistant by including a `metadata` field in your request, or update it later using the update assistant endpoint.

  For more information, see [Create an assistant](/reference/api/latest/assistant/create_assistant) and [Manage assistants](/guides/assistant/manage-assistants).
</Update>

<Update label="2025-12-23" tags={["Docs"]}>
  ### Test Pinecone at scale

  Added a new guide to help you [test Pinecone at production scale](/guides/get-started/test-at-scale). The guide describes how to create an index, import 10 million vectors, and then use Vector Search Bench (VSB) to capture performance metrics for 100,000 queries. To run this test, consider signing up for a [Standard plan trial](/guides/organizations/manage-billing/standard-trial).
</Update>

<Update label="2025-12-22" tags={["Assistant"]}>
  ### Pinecone Assistant now supports GPT-5

  Pinecone Assistant now supports the GPT-5 model. To use it, set `model` to `gpt-5` when [chatting with your assistant](/reference/api/latest/assistant/chat_assistant).

  For more information, see [Chat with Assistant](/guides/assistant/chat-with-assistant#choose-a-model) and [Chat through the OpenAI-compatible interface](/guides/assistant/chat-through-the-openai-compatible-interface#choose-a-model).
</Update>

<Update label="2025-12-12" tags={["General"]}>
  ### Upgrade from Starter to Standard trial

  Organizations on the Starter plan can now upgrade to a Standard plan trial at any time. The Standard trial provides 21 days and \$300 in credits to test Pinecone at scale, with access to Standard plan features such as higher limits. For more information, see [Standard plan trial](/guides/organizations/manage-billing/standard-trial).
</Update>

<Update label="2025-12-07" tags={["Assistant"]}>
  ### Delete Assistant files while they are still processing

  You can now delete a file uploaded to an assistant while it is still processing. Previously, you had to wait for file processing to complete before you could delete a file.

  For more information, see [Manage files](/guides/assistant/manage-files).
</Update>

<Update label="2025-12-02" tags={["General"]}>
  ### Annual commit discount

  In the Pinecone console, customers on Standard and Enterprise pay-as-you-go plans can now commit to an annual contract and receive a discount on usage. For more information, see [Understanding cost](/guides/manage-cost/understanding-cost#discounts).
</Update>

<Update label="2025-12-02" tags={["Assistant"]}>
  ### Increased instructions limit for assistants

  The maximum size for assistant instructions has been increased from 8 KB to 16 KB.

  Instructions are included in every chat API call. Longer instructions increase input token costs for each request and consume more of the LLM's context window, reducing available space for retrieved context and conversation history.

  For more information, see [Create an assistant](/reference/api/latest/assistant/create_assistant) and [Manage assistants](/guides/assistant/manage-assistants).
</Update>

<Update label="2025-12-01" tags={["API"]}>
  ### Dedicated Read Nodes: now in public preview

  [Dedicated Read Nodes](/guides/index-data/dedicated-read-nodes) is now in [public preview](/release-notes/feature-availability). Users on Standard and Enterprise plans can now create Dedicated Read Nodes indexes using the console and API.

  With Dedicated Read Nodes, you can provision read hardware for large, high-throughput indexes that require predictable, low latency.

  For more information, see [Dedicated Read Nodes](/guides/index-data/dedicated-read-nodes).
</Update>

## November 2025

<Update label="2025-11-19" tags={["API"]}>
  ### Pinecone API version `2025-10` is now the latest stable version

  `2025-10` is now the latest [stable version](/reference/api/versioning#release-schedule) of the [Pinecone APIs](/reference/api/introduction). This version does not include any breaking changes, and provides the following new features:

  **[Dedicated read nodes](/guides/index-data/dedicated-read-nodes) (early access):**

  * The [Create an index](/reference/api/2025-10/control-plane/create_index) and [Configure index](/reference/api/2025-10/control-plane/configure_index) endpoints allow you to create and configure indexes that use dedicated read nodes. With dedicated read nodes, you can allocate dedicated hardware for read operations. This is useful for large, high-QPS indexes that require consistent, predictable low latency.
  * The [Describe an index](/reference/api/2025-10/control-plane/describe_index) endpoint now allows you to view information about a dedicated index: shards, replicas, and scaling status.

  **Namespace and metadata schema management:**

  * The [Create a namespace](/reference/api/2025-10/data-plane/createnamespace) endpoint allows you to create a namespace without upserting vectors. You can optionally configure a metadata schema when creating the namespace to pre-declare which metadata fields should be indexed for filtering.
  * The [List namespaces](/reference/api/2025-10/data-plane/listnamespaces) endpoint now supports filtering namespaces by prefix and returns the total count of matching namespaces.
  * The [Describe a namespace](/reference/api/2025-10/data-plane/describenamespace) endpoint allows you to view the metadata schema configuration for a namespace, including which fields are indexed for filtering.
  * The [Create an index](/reference/api/2025-10/control-plane/create_index) endpoint now allows you to specify a metadata schema at index creation time to pre-declare which metadata fields should be indexed for filtering across all namespaces.

  **Update and fetch by metadata:**

  * The [Update a vector](/reference/api/2025-10/data-plane/update) endpoint allows you to update metadata across multiple records in a namespace using a metadata filter expression, eliminating the need to update records individually by ID.
  * The [Fetch vectors by metadata](/reference/api/2025-10/data-plane/fetch_by_metadata) endpoint allows you to fetch vectors using metadata filters without knowing their vector IDs.

  **Enhanced sparse search:**

  * The [Search with text](/reference/api/2025-10/data-plane/search_records) endpoint now supports a `match_terms` parameter that allows you to specify terms that must be present in search results for sparse indexes.
</Update>

<Update label="2025-11-03" tags={["Assistant"]}>
  ### n8n quickstarts

  Added new quickstart options to create an n8n workflow that downloads files via HTTP and lets you chat with them using Pinecone and OpenAI.

  * [n8n quickstart for Pinecone Assistant](/guides/assistant/quickstart/n8n-quickstart)
  * [n8n quickstart for Pinecone Database](/guides/get-started/quickstart#n8n)
</Update>

<Update label="2025-11-03" tags={["SDK"]}>
  ### Released Node.js SDK v6.1.3

  Released [`v6.1.3`](https://github.com/pinecone-io/pinecone-ts-client/releases/tag/v6.1.3) of the [Pinecone Node.js SDK](/reference/sdks/node/overview).

  This version of the SDK fixes a bug in `Assistant.listFiles()`. Previously, when passing a `filter` to `listFiles`, the top-level `metadata` object was not handled correctly. This caused the method to return all files, regardless of the filter.

  It's no longer necessary to provide a top-level `metadata` object. Instead, declare metadata fields directly in the `filter` object:

  * ✅ `const files = await assistant.listFiles({ filter: { document_type: 'manuscript' } });`
  * ❌ `const files = await assistant.listFiles({ filter: { metadata: { document_type: 'manuscript' } } });`

  If you're using the old syntax, update it so that your filter works correctly. For more information about listing the files associated with an assistant, see [Manage files](/guides/assistant/manage-files).
</Update>

## October 2025

<Update label="2025-10-30" tags={["Assistant"]}>
  ### Increased files per assistant on the Starter plan

  On the Starter plan, you can now upload up to 100 files to an assistant. Previously, the limit was 10 files.

  To learn more, see [Assistant limits](/guides/assistant/pricing-and-limits#assistant-limits).
</Update>

<Update label="2025-10-23" tags={["General"]}>
  ### Enhanced monthly spend alerts

  You can now set multiple spend alerts to monitor your organization's monthly spending. These alerts notify designated recipients when spending reaches specified thresholds. The alerts automatically reset at the start of each monthly billing cycle.

  Additionally, to protect from unexpected cost increases, Pinecone sends an alert when spending exceeds double your previous month's invoice amount. While the alert threshold is fixed, you can modify which email addresses receive the alert and enable or disable the alert notifications.

  To learn more, see [Manage cost](/guides/manage-cost/manage-cost).
</Update>

<Update label="2025-10-16" tags={["Docs"]}>
  ### Agentic quickstart

  Added new [agentic quickstart](/guides/get-started/quickstart#cursor) options to help you build Pinecone applications with AI coding agents like Claude Code and Cursor. Instead of copying code snippets, you work with an agent that understands Pinecone APIs and implements production-ready patterns automatically.
</Update>

<Update label="2025-10-16" tags={["Docs"]}>
  ### AI Engine integration

  Added the [AI Engine](/integrations/ai-engine) integration page.
</Update>

<Update label="2025-10-09" tags={["CLI"]}>
  ### Pinecone CLI v0.1.0

  We've released [v0.1.0](https://github.com/pinecone-io/cli/releases/tag/v0.1.0) of the [Pinecone CLI](https://github.com/pinecone-io/cli). The CLI lets you manage Pinecone infrastructure (organizations, projects, indexes, and API keys) directly from your terminal and in CI/CD.

  This feature is in [public preview](/release-notes/feature-availability). We'll be adding more features to the CLI over time, and we'd love your [feedback](https://community.pinecone.io/) on this early version.

  For more information, see the [CLI quickstart](/reference/cli/quickstart).
</Update>

## September 2025

<Update label="2025-09-30" tags={["Docs"]}>
  ### Production best practices

  Added a new [error handling guide](/guides/production/error-handling) to help you handle errors gracefully in production, including implementing retry logic with exponential backoff for rate limits and transient errors.

  Updated the [production checklist](/guides/production/production-checklist) with enhanced guidance on data modeling, database limits, and performance optimization.
</Update>

<Update label="2025-09-11" tags={["Docs"]}>
  ### Changing payment methods

  Added a new guide to help customers [change their payment method](/guides/organizations/manage-billing/change-payment-method) for Pinecone's Standard or Enterprise plan, including switching from credit card to marketplace billing and vice versa.
</Update>

<Update label="2025-09-04" tags={["Database"]}>
  ### Released Pinecone Terraform Provider v2.0.0

  Released [v2.0.0](https://github.com/pinecone-io/terraform-provider-pinecone/releases/tag/v2.0.0) of the [Terraform Provider for Pinecone](/integrations/terraform). This version adds support for managing API keys and projects.
</Update>

<Update label="2025-09-02" tags={["Assistant"]}>
  ### Multimodal context for assistants

  Assistants can now gather context from images in PDF files. To learn more, see [Multimodal context for assistants](/guides/assistant/multimodal). This feature is in [public preview](/release-notes/feature-availability).
</Update>

## August 2025

<Update label="2025-08-29" tags={["Database"]}>
  ### Filter lexical search by required terms

  You can now filter lexical search results to require specific terms. This is especially useful for filtering out results that don't contain essential keywords, requiring domain-specific terminology in results, and ensuring specific people, places, or things are mentioned. This feature is in [public preview](/release-notes/feature-availability).

  To learn more, see [Filter by required terms](/guides/search/lexical-search#filter-by-required-terms).
</Update>

<Update label="2025-08-26" tags={["Docs"]}>
  ### Zapier integration

  Added the [Zapier](/integrations/zapier) integration page.
</Update>

<Update label="2025-08-25" tags={["General"]}>
  ### SSO setup improvements

  We've streamlined the SSO setup process, eliminating the need to add placeholder URLs to your identity provider. To learn more, see [Configure SSO with Okta](/guides/production/configure-single-sign-on/okta).
</Update>

<Update label="2025-08-25" tags={["Database"]}>
  ### Update metadata across multiple records

  You can now [update metadata across multiple records](/guides/manage-data/update-data#update-metadata-across-multiple-records) in a namespace. This feature is in [early access](/release-notes/feature-availability).
</Update>

<Update label="2025-08-13" tags={["Database"]}>
  ### Data import from Azure Blob Storage

  Now, you can import data from an Azure Blob Storage container into a Pinecone index. This feature is in [public preview](/release-notes/feature-availability).

  To learn more, read:

  * [Integrate with Azure Blob Storage](/guides/operations/integrations/integrate-with-azure-blob-storage)
  * [Import records](/guides/index-data/import-data)
  * [Pinecone's pricing](https://www.pinecone.io/pricing/)
</Update>

<Update label="2025-08-12" tags={["Assistant"]}>
  ### Assistant MCP server endpoint update

  **Breaking Change**: After August 31, 2025 at 11:59:59 PM UTC, the SSE-based MCP endpoint for assistants (`/mcp/assistants/<YOUR_ASSISTANT_NAME>/sse`) will no longer work.

  Before then, update your applications to use the streamable HTTP transport MCP endpoint (`/mcp/assistants/<YOUR_ASSISTANT_NAME>`). This endpoint follows the current [MCP protocol specification](https://modelcontextprotocol.io/specification/2025-06-18/basic/transports#streamable-http) and provides improved flexibility and compatibility.

  Please note that Assistant MCP servers are in [early access](/release-notes/feature-availability) and are not intended for production usage.

  For more information, see [Use an Assistant MCP server](/guides/assistant/mcp-server).
</Update>

<Update label="2025-08-04" tags={["Docs"]}>
  ### VoltAgent integration

  Added the [VoltAgent](/integrations/voltagent) integration page.
</Update>

## July 2025

<Update label="2025-07-28" tags={["Database"]}>
  ### Increased context window for `pinecone-sparse-english-v0`

  You can now raise the context window for Pinecone's hosted [`pinecone-sparse-english-v0`](/guides/index-data/create-an-index#pinecone-sparse-english-v0) embedding model from `512` to `2048` using the `max_tokens_per_sequence` parameter.
</Update>

<Update label="2025-07-23" tags={["SDK"]}>
  ### Release Go SDK v4.1.0, v4.1.1, and v4.1.2

  Released [`v4.1.0`](https://github.com/pinecone-io/go-pinecone/releases/tag/v4.1.0), [`v4.1.1`](https://github.com/pinecone-io/go-pinecone/releases/tag/v4.1.1), and [`v4.1.2`](https://github.com/pinecone-io/go-pinecone/releases/tag/v4.1.2) of the [Pinecone Go SDK](/reference/sdks/go/overview).

  * `v4.1.0` adds support for admin API operations for working with API keys, projects, and service accounts.
  * `v4.1.1` adds `PercentComplete` and `RecordsImported` to the response when [describing an import](/guides/index-data/import-data#track-import-progress) and [listing imports](/guides/index-data/import-data#list-imports).
  * `v4.1.2` adds support for [migrating a pod-based index to serverless](/guides/indexes/pods/migrate-a-pod-based-index-to-serverless#3-start-migration).
</Update>

<Update label="2025-07-23" tags={["SDK"]}>
  ### Release Node.js SDK v6.1.2

  Released [`v6.1.2`](https://github.com/pinecone-io/pinecone-ts-client/releases/tag/v6.1.2) of the [Pinecone Node.js SDK](/reference/sdks/node/overview). This version adds support for the following:

  * [Migrating a pod-based index to serverless](/guides/indexes/pods/migrate-a-pod-based-index-to-serverless#3-start-migration).
  * Controlling whether `signed_url` is included in the response when [describing a file](/guides/assistant/manage-files#get-the-status-of-a-file) for an assistant.
</Update>

## June 2025

<Update label="2025-06-26" tags={["Assistant"]}>
  ### Unlimited assistant file storage for paid plans

  Organizations on the [Standard and Enterprise plans](https://www.pinecone.io/pricing/) now have [unlimited file storage](/reference/api/assistant/assistant-limits) for their assistants. Previously, organizations on these plans were limited to 10 GB of file storage per project.
</Update>

<Update label="2025-06-23" tags={["Database"]}>
  ### Data import from Google Cloud Storage

  You can now [import data](/guides/index-data/import-data) into an index from [Google Cloud Storage](/guides/operations/integrations/integrate-with-google-cloud-storage). This feature is in [public preview](/release-notes/feature-availability).
</Update>

<Update label="2025-06-20" tags={["SDK"]}>
  ### Released Python SDK v7.1.0, v7.2.0, and v7.3.0

  Released [`v7.1.0`](https://github.com/pinecone-io/pinecone-python-client/releases/tag/v7.1.0), [`v7.2.0`](https://github.com/pinecone-io/pinecone-python-client/releases/tag/v7.2.0), and [`v7.3.0`](https://github.com/pinecone-io/pinecone-python-client/releases/tag/v7.3.0) of the [Pinecone Python SDK](/reference/sdks/python/overview).

  * `v7.1.0` fixes minor bugs.
  * `v7.2.0` adds support for [managing namespaces](/guides/manage-data/manage-namespaces).
  * `v7.3.0` adds support for admin API operations for working with API keys, projects, and service accounts.
</Update>

<Update label="2025-06-16" tags={["SDK"]}>
  ### Released Go SDK v4.0.0 and v4.0.1

  Released [`v4.0.0`](https://github.com/pinecone-io/go-pinecone/releases/tag/v4.0.0) and [`v4.0.1`](https://github.com/pinecone-io/go-pinecone/releases/tag/v4.0.1) of the [Pinecone Go SDK](/reference/sdks/go/overview).

  Go SDK `v4.0.0` uses the latest stable API version, `2025-04`, and includes support for the following:

  * [Managing namespaces](/guides/manage-data/manage-namespaces)
  * [Reusing an index connection with a new namespace](/guides/manage-data/target-an-index#target-by-index-host-recommended) (see the Go example)
  * [Creating and managing backups](/guides/manage-data/back-up-an-index)
  * [Restoring indexes from backups](/guides/manage-data/restore-an-index)
  * [Listing embedding and reranking models hosted by Pinecone](/reference/api/2025-04/inference/list_models)
  * [Getting details about a model hosted by Pinecone](/reference/api/2025-04/inference/describe_model)

  Go SDK `v4.0.1` expands the [`DescribeIndex`](/guides/production/configure-private-endpoints#read-and-write-data) response to include the `private_host` value for connecting to indexes with a private endpoint.
</Update>

<Update label="2025-06-16" tags={["SDK"]}>
  ### Released Node.js SDK v6.1.1

  Released [`v6.1.1`](https://github.com/pinecone-io/pinecone-ts-client/releases/tag/v6.1.1) of the [Pinecone Node.js SDK](/reference/sdks/node/overview). This version adds support for [setting the sampling temperature](/guides/assistant/chat-with-assistant#set-the-sampling-temperature) for an assistant, and expands the [`describeIndex`](/guides/production/configure-private-endpoints#read-and-write-data) response to include the `private_host` value for connecting to indexes with a private endpoint.
</Update>

<Update label="2025-06-12" tags={["Docs"]}>
  ### Data modeling guide

  Added a new guide to help you [model your data](/guides/index-data/data-modeling) for efficient ingestion, retrieval, and management in Pinecone.
</Update>

<Update label="2025-06-05" tags={["SDK"]}>
  ### Released Java SDK v5.1.0

  Released [`v5.1.0`](https://github.com/pinecone-io/pinecone-java-client/releases/tag/v5.1.0) of the [Pinecone Java SDK](/reference/sdks/java/overview). This version adds support for [listing](/reference/api/2025-04/inference/list_models) and [describing](/reference/api/2025-04/inference/describe_model) embedding and reranking models hosted by Pinecone.
</Update>

<Update label="2025-06-05" tags={["SDK"]}>
  ### Released Node.js SDK v6.1.0

  Released [`v6.1.0`](https://github.com/pinecone-io/pinecone-ts-client/releases/tag/v6.1.0) of the [Pinecone Node.js SDK](/reference/sdks/node/overview). This version adds support for [controlling the context snippets sent to the LLM](/guides/assistant/chat-with-assistant#control-the-context-snippets-sent-to-the-llm) by an assistant.
</Update>

## May 2025

<Update label="2025-05-29" tags={["SDK"]}>
  ### Released Python SDK v7.0.1 and v7.0.2

  Released [`v7.0.1`](https://github.com/pinecone-io/pinecone-python-client/releases/tag/v7.0.1) and [`v7.0.2`](https://github.com/pinecone-io/pinecone-python-client/releases/tag/v7.0.2) of the [Pinecone Python SDK](/reference/sdks/python/overview). These versions fix minor bugs discovered since the release of the `v7.0.0` major version.
</Update>

<Update label="2025-05-29" tags={["SDK"]}>
  ### Released Node.s SDK v6.0.1

  Released [`v6.0.1`](https://github.com/pinecone-io/pinecone-ts-client/releases/tag/v6.0.1) of the [Pinecone Node.js SDK](/reference/sdks/node/overview). This version adds pagination to the [`listBackups`](/guides/manage-data/back-up-an-index#list-backups-in-a-project) operation.
</Update>

<Update label="2025-05-19" tags={["API"]}>
  ### Pinecone API version `2025-04` is now the latest stable version

  `2025-04` is now the latest [stable version](/reference/api/versioning#release-schedule) of the [Pinecone APIs](/reference/api/introduction). For highlights, see the SDK releases below.
</Update>

<Update label="2025-05-19" tags={["SDK"]}>
  ### Released Python SDK v7.0.0

  Released [`v7.0.0`](https://github.com/pinecone-io/pinecone-python-client/releases/tag/v7.0.0) of the [Pinecone Python SDK](/reference/sdks/python/overview). This version uses the latest stable API version, `2025-04`, and includes support for the following:

  * [Creating and managing backups](/guides/manage-data/back-up-an-index)
  * [Restoring indexes from backups](/guides/manage-data/restore-an-index)
  * [Listing embedding and reranking models hosted by Pinecone](/reference/api/2025-04/inference/list_models)
  * [Getting details about a model hosted by Pinecone](/reference/api/2025-04/inference/describe_model)
  * [Creating a BYOC index](/guides/production/bring-your-own-cloud#create-an-index)

  Additionally, the `pinecone-plugin-assistant` package required to work with [Pinecone Assistant](/guides/assistant/overview) is now included by default; it is no longer necessary to install the plugin separately.
</Update>

<Update label="2025-05-19" tags={["SDK"]}>
  ### Released Node.js SDK v6.0.0

  Released [`v6.0.0`](https://github.com/pinecone-io/pinecone-ts-client/releases/tag/v6.0.0) of the [Pinecone Node.js SDK](/reference/sdks/node/overview). This version uses the latest stable API version, `2025-04`, and includes support for the following:

  * [Managing namespaces](/guides/manage-data/manage-namespaces)
  * [Creating and managing backups](/guides/manage-data/back-up-an-index)
  * [Restoring indexes from backups](/guides/manage-data/restore-an-index)
  * [Listing embedding and reranking models hosted by Pinecone](/reference/api/2025-04/inference/list_models)
  * [Getting details about a model hosted by Pinecone](/reference/api/2025-04/inference/describe_model)
</Update>

<Update label="2025-05-19" tags={["SDK"]}>
  ### Released Java SDK v5.0.0

  Released [`v5.0.0`](https://github.com/pinecone-io/pinecone-java-client/releases/tag/v5.0.0) of the [Pinecone Java SDK](/reference/sdks/java/overview). This version uses the latest stable API version, `2025-04`, and includes support for the following:

  * [Creating indexes with integrated embedding and reranking](/guides/index-data/indexing-overview#integrated-embedding)
  * [Upserting text to an integrated index](/guides/index-data/upsert-data)
  * [Searching an integrated index with text](/guides/search/semantic-search#search-with-text)
  * [Managing namespaces](/guides/manage-data/manage-namespaces)
  * [Creating and managing backups](/guides/manage-data/back-up-an-index)
  * [Restoring indexes from backups](/guides/manage-data/restore-an-index)
</Update>

<Update label="2025-05-19" tags={["SDK"]}>
  ### Released .NET SDK v4.0.0

  Released [`v4.0.0`](https://github.com/pinecone-io/pinecone-dotnet-client/releases/tag/4.0.0) of the [Pinecone .NET SDK](/reference/sdks/dotnet/overview). This version uses the latest stable API version, `2025-04`, and includes support for the following:

  * [Creating indexes with integrated embedding and reranking](/guides/index-data/indexing-overview#integrated-embedding)
  * [Upserting text to an integrated index](/guides/index-data/upsert-data)
  * [Searching an integrated index with text](/guides/search/semantic-search#search-with-text)
  * [Managing namespaces](/guides/manage-data/manage-namespaces)
  * [Creating and managing backups](/guides/manage-data/back-up-an-index)
  * [Restoring indexes from backups](/guides/manage-data/restore-an-index)
  * [Listing embedding and reranking models hosted by Pinecone](/reference/api/2025-04/inference/list_models)
  * [Getting details about a model hosted by Pinecone](/reference/api/2025-04/inference/describe_model)

  <Warning>
    Before upgrading to `v4.0.0`, update all relevant code to account for the following [breaking changes](/reference/api/versioning#breaking-changes). See the [`v4.0.0`](https://github.com/pinecone-io/pinecone-dotnet-client/releases/tag/4.0.0) release notes for full details.

    * The [`create_index`](/reference/api/2025-04/control-plane/create_index) and [`create_for_model`](/reference/api/2025-04/control-plane/create_for_model) operations:
      * `CreateIndexRequestMetric` has been renamed to `MetricType`.
    * The [`list_indexes`](/reference/api/2025-04/control-plane/list_indexes) operation:
      * `ModelIndexEmbedMetric` has been renamed to `MetricType`.
    * The [`embed`](/reference/api/2025-04/inference/generate-embeddings) operation:
      * `SparseEmbedding.SparseIndices` has changed from `IEnumerable<int>` to `IEnumerable<long>`.
  </Warning>
</Update>

<Update label="2025-05-06" tags={["Docs"]}>
  ### New Docs IA

  We've overhauled the information architecture of our guides to mirror the goals of users, from indexing to searching to optimizing to production.

  This change includes distinct pages for search types:

  * [Semantic search](https://docs.pinecone.io/guides/search/semantic-search)
  * [Lexical search](https://docs.pinecone.io/guides/search/lexical-search)
  * [Hybrid search](https://docs.pinecone.io/guides/search/hybrid-search)

  And optimization techniques:

  * [Increase relevance](https://docs.pinecone.io/guides/optimize/increase-relevance)
  * [Increase throughput](https://docs.pinecone.io/guides/optimize/increase-throughput)
  * [Decrease latency](https://docs.pinecone.io/guides/optimize/decrease-latency)
</Update>

## April 2025

<Update label="2025-04-25" tags={["Database"]}>
  ### Bring Your Own Cloud (BYOC) in GCP

  The [Bring Your Own Cloud (BYOC)](/guides/production/bring-your-own-cloud) offering is now available in GCP. Organizations with high security and compliance requirements can use BYOC to deploy Pinecone Database in their own GCP account. This feature is in [public preview](/release-notes/feature-availability).
</Update>

<Update label="2025-04-21" tags={["Database"]}>
  ### Integrate AI agents with Pinecone MCP

  [Pinecone's open-source MCP server](/guides/operations/mcp-server) enables AI agents to interact directly with Pinecone's functionality and documentation via the standardized [Model Context Protocol (MCP)](https://modelcontextprotocol.io/l). Using the MCP server, agents can search Pinecone documentation, manage indexes, upsert data, and query indexes for relevant information.
</Update>

<Update label="2025-04-21" tags={["Assistant"]}>
  ### Add context to AI agents with Assistant MCP

  Every Pinecone Assistant now has a [dedicated MCP server](/guides/assistant/mcp-server) that gives AI agents direct access to the assistant's knowledge through the standardized [Model Context Protocol (MCP)](https://modelcontextprotocol.io/).
</Update>

<Update label="2025-04-21" tags={["Assistant"]}>
  ### Upload a file from an in-memory binary stream

  You can [upload a file to an assistant directly from an in-memory binary stream](/guides/assistant/upload-files#upload-from-a-binary-stream) using the Python SDK and the BytesIO class.
</Update>

<Update label="2025-04-21" tags={["Database"]}>
  ### Released Pinecone Terraform Provider v1.0.0

  Released [v1.0.0](https://github.com/pinecone-io/terraform-provider-pinecone/releases/tag/v1.0.0) of the [Terraform Provider for Pinecone](/integrations/terraform). This version adds support for [sparse indexes](/guides/index-data/indexing-overview#indexes-with-sparse-vectors), [indexes with integrated embedding and reranking](/guides/index-data/indexing-overview#integrated-embedding), [index tags](/guides/manage-data/manage-indexes#configure-index-tags), and [index deletion protection](/guides/manage-data/manage-indexes#configure-deletion-protection).
</Update>

<Update label="2025-04-13" tags={["SDK"]}>
  ### Released .NET SDK v3.1.0

  Released [`v3.1.0`](https://github.com/pinecone-io/pinecone-dotnet-client/releases/tag/3.1.0) of the [Pinecone .NET SDK](/reference/sdks/dotnet/overview). This version adds support for [indexes with integrated embedding and reranking](/guides/index-data/indexing-overview#integrated-embedding).
</Update>

<Update label="2025-04-04" tags={["Docs"]}>
  ### LLM shortcuts for Pinecone docs

  You can now use the "Copy page" options at the top of every page of the Pinecone documentation to quickly ground LLMs with Pinecone-specific context.
</Update>

## March 2025

<Update label="2025-03-26" tags={["Assistant"]}>
  ### Control the context snippets the assistant sends to the LLM

  You can [control the context snippets sent to the LLM](/guides/assistant/chat-with-assistant#control-the-context-snippets-sent-to-the-llm) by setting `context_options` in the request.
</Update>

<Update label="2025-03-24" tags={["SDK"]}>
  ### Released Go SDK v3.1.0

  Released [`v3.1.0`](https://github.com/pinecone-io/go-pinecone/releases/tag/v3.1.0) of the [Pinecone Go SDK](/reference/sdks/go/overview). This version adds support for [indexes with integrated embedding and reranking](/guides/index-data/indexing-overview#integrated-embedding).
</Update>

<Update label="2025-03-21" tags={["General"]}>
  <img className="block max-w-full" noZoom src="https://mintcdn.com/pinecone/r0TaYXrfSrAYZYUj/images/release-notes/launch-week-march-2025.png?fit=max&auto=format&n=r0TaYXrfSrAYZYUj&q=85&s=a5a2d145f08f688f6ae41242a0b5ff45" width="2400" height="1260" data-path="images/release-notes/launch-week-march-2025.png" />

  ### Launch week: Dark mode

  Dark mode is now out for Pinecone's website, docs, and console. You can change your theme at the top right of each site.
</Update>

<Update label="2025-03-20" tags={["General"]}>
  ### Launch week: Self-service audit logs

  You can now enable and [configure audit logs](/guides/production/configure-audit-logs) for your Pinecone organization. [Audit logs](/guides/production/security-overview#audit-logs) provide a detailed record of user, service account, and API actions that occur within Pinecone. This feature is in [public preview](/release-notes/feature-availability) and available only on [Enterprise plans](https://www.pinecone.io/pricing/).
</Update>

<Update label="2025-03-19" tags={["General"]}>
  ### Launch week: Introducing the Admin API and service accounts

  You can now use [service accounts](/guides/organizations/understanding-organizations#service-accounts) to programmatically manage your Pinecone organization through the Admin API. Use the Admin API to [create](/guides/projects/create-a-project) and [manage projects](/guides/projects/manage-projects), as well as [create and manage API keys](/guides/projects/manage-api-keys). The Admin API and service accounts are in [public preview](/release-notes/feature-availability).
</Update>

<Update label="2025-03-18" tags={["Database"]}>
  ### Launch week: Back up an index through the API

  You can now [back up an index](/guides/manage-data/back-up-an-index) and [restore an index](/guides/manage-data/restore-an-index) through the Pinecone API. This feature is in [public preview](/release-notes/feature-availability).
</Update>

<Update label="2025-03-17" tags={["Database"]}>
  ### Launch week: Optimized database architecture

  Pinecone has optimized its [serverless database architecture](/guides/get-started/database-architecture) to meet the growing demand for large-scale agentic workloads and improved performance for search and recommendation workloads. New customers will use this architecture by default, and existing customers will gain access over the next month.
</Update>

<Update label="2025-03-13" tags={["Docs"]}>
  ### Firebase Genkit integration

  Added the [Firebase Genkit](/integrations/genkit) integration page.
</Update>

<Update label="2025-03-10" tags={["Database"]}>
  ### Bring Your Own Cloud (BYOC) in public preview

  [Bring Your Own Cloud (BYOC)](/guides/production/bring-your-own-cloud) lets you deploy Pinecone Database in your private AWS account to ensure data sovereignty and compliance, with Pinecone handling provisioning, operations, and maintenance. This feature is in [public preview](/release-notes/feature-availability) on AWS.
</Update>

## February 2025

<Update label="2025-02-21" tags={["Docs"]}>
  ### Docs site refresh

  We've refreshed the look and layout of the [Pinecone documentation](https://docs.pinecone.io) site. You can now use the dropdown at the top of the side navigation to view documentation for either [Pinecone Database](/guides/get-started/overview) or [Pinecone Assistant](/guides/assistant/overview).
</Update>

<Update label="2025-02-18" tags={["Assistant"]}>
  ### Limit the number of chunks retrieved

  You can now limit the number of chunks the reranker sends to the LLM. To do this, set the `top_k` parameter (default is 15) when [retrieving context snippets](/guides/assistant/retrieve-context-snippets).
</Update>

<Update label="2025-02-21" tags={["Docs"]}>
  ### Assistant Quickstart colab notebook

  Added the [Assistant Quickstart colab notebook](https://colab.research.google.com/github/pinecone-io/examples/blob/master/docs/assistant-quickstart.ipynb). This notebook shows you how to set up and use [Pinecone Assistant](/guides/assistant/overview) in your browser.
</Update>

<Update label="2025-02-18" tags={["SDK"]}>
  ### Released Node.js SDK v5.0.0

  Released [`v5.0.0`](https://github.com/pinecone-io/pinecone-ts-client/releases/tag/v5.0.0) of the [Pinecone Node.js SDK](/reference/sdks/node/overview). This version uses the latest stable API version, `2025-01`, and includes support for [Pinecone Assistant](/guides/assistant/overview) and [sparse-only indexes](/guides/index-data/indexing-overview#indexes-with-sparse-vectors).
</Update>

<Update label="2025-02-11" tags={["Docs"]}>
  ### New integrations

  Added the [Box](/integrations/box) and [Cloudera AI](/integrations/cloudera) integration pages.
</Update>

<Update label="2025-02-10" tags={["Assistant"]}>
  ### Citation highlights in assistant responses

  You can now include [highlights](/guides/assistant/chat-with-assistant#include-citation-highlights-in-the-response) in an assistant's citations. Highlights are the specific parts of the document that the assistant used to generate the response.

  Citation highlights are available in the Pinecone console or API versions `2025-04` and later.
</Update>

<Update label="2025-02-07" tags={["API"]}>
  ### Pinecone API version `2025-01` is now the latest stable version

  `2025-01` is now the latest [stable version](/reference/api/versioning#release-schedule) of the [Pinecone APIs](/reference/api/introduction).
</Update>

<Update label="2025-02-07" tags={["SDK"]}>
  ### Released Python SDK v6.0.0

  Released [`v6.0.0`](https://github.com/pinecone-io/pinecone-python-client/releases/tag/v6.0.0) of the [Pinecone Python SDK](/reference/sdks/python/overview). This version uses the latest stable API version, `2025-01`, and includes support for the following:

  * [Index tags](/guides/manage-data/manage-indexes#configure-index-tags) to categorize and identify your indexes.
  * [Integrated inference](/reference/api/introduction#inference) without the need for extra plugins. If you were using the preview functionality of integrated inference, you must uninstall the `pinecone-plugin-records` package to use the `v6.0.0` release.
  * Enum objects to help with the discoverability of some configuration options, for example, `Metric`, `AwsRegion`, `GcpRegion`, `PodType`, `EmbedModel`, `RerankModel`. This is a backwards compatible change; you can still pass string values for affected fields.
  * New client variants, `PineconeAsyncio` and `IndexAsyncio`, which provide `async` methods for use with [asyncio](https://docs.python.org/3/library/asyncio.html). This makes it possible to use Pinecone with modern async web frameworks such as [FastAPI](https://fastapi.tiangolo.com/), [Quart](https://quart.palletsprojects.com/en/latest/), and [Sanic](https://sanic.dev/en/). Async support should significantly increase the efficiency of running many upserts in parallel.

  <Warning>
    Before upgrading to `v6.0.0`, update all relevant code to account for the following [breaking changes](/reference/api/versioning#breaking-changes). See the [`v6.0.0`](https://github.com/pinecone-io/pinecone-python-client/releases/tag/v6.0.0) release notes for full details.

    * Incorporated the `pinecone-plugin-records` and `pinecone-plugin-inference` plugins into the `pinecone` package. If you are using these plugins, you must unstall them to use `v6.0.0`.
    * Dropped support for Python 3.8, which has now reached official end of life, and added support for Python 3.13.
    * Removed the explicit dependency on `tqdm`, which is used to provide a progress bar when upserting data into Pinecone. If `tqdm` is available in the environment, the Pinecone SDK will detect and use it, but `tdqm` is no longer required to run the SDK. Popular notebook platforms such as [Jupyter](https://jupyter.org/) and [Google Colab](https://colab.google/) already include `tqdm` in the environment by default, but if you are running small scripts in other environments and want to continue seeing progress bars, you will need to separately install the `tqdm` package.
    * Removed some previously deprecated and rarely used keyword arguments (`config`, `openapi_config`, and `index_api`) to instead prefer dedicated keyword arguments for individual settings such as `api_key`, `proxy_url`, etc.
  </Warning>
</Update>

<Update label="2025-02-07" tags={["SDK"]}>
  ### Released Java SDK v4.0.0

  Released [`v4.0.0`](https://github.com/pinecone-io/pinecone-java-client/releases/tag/v4.0.0) of the [Pinecone Java SDK](/reference/sdks/java/overview). This version uses the latest stable API version, `2025-01`, and adds support for [sparse-only indexes](/guides/index-data/indexing-overview#indexes-with-sparse-vectors).

  <Warning>
    Before upgrading to `v4.0.0`, update all relevant code to account for the following [breaking changes](/reference/api/versioning#breaking-changes). See the [`v4.0.0`](https://github.com/pinecone-io/pinecone-java-client/releases/tag/v4.0.0) release notes for full details.

    * [`embed` method](/reference/api/2025-01/inference/generate-embeddings):
      * `parameters` now accepts `Map<String, Object>` instead of `EmbedRequestParameters`.
      * The `Embeddings` response class now has dense and sparse embeddings. You now must use `getDenseEmbedding()` or `getSparseEmbedding()`. For example, instead of `embeddings.getData().get(0).getValues()`, you would use `embeddings.getData().get(0).getDenseEmbedding().getValues()`.

    * [`rerank` method](/guides/search/rerank-results):
      * `documents` now accepts `List<Map<String, Object>>` instead of `List<Map<String, String>>`.
      * `parameters` now accepts `Map<String, Object>` instead of `Map<String, String>`.
  </Warning>
</Update>

<Update label="2025-02-07" tags={["SDK"]}>
  ### Released Go SDK v3.0.0

  Released [`v3.0.0`](https://github.com/pinecone-io/go-pinecone/releases/tag/v3.0.0) of the [Pinecone Go SDK](/reference/sdks/go/overview). This version uses the latest stable API version, `2025-01`, and adds support for [sparse-only indexes](/guides/index-data/indexing-overview#indexes-with-sparse-vectors).

  <Warning>
    Before upgrading to `v3.0.0`, update all relevant code to account for the following [breaking changes](/reference/api/versioning#breaking-changes). See the [`v3.0.0`](https://github.com/pinecone-io/go-pinecone/releases/tag/v3.0.0) release notes for full details.

    * [`embed` operation](/reference/api/2025-01/inference/generate-embeddings):
      * `EmbedParameters` is no longer typed as a pointer.
    * [`create_index` operation](/guides/index-data/create-an-index):
      * `CreateServerlessIndexRequest` and `CreatePodIndexRequest` structs have been updated, and fields are now classified as pointers to better denote optionality around creating specific types of indexes: `Metric`, `Dimension`, `VectorType`, and `DeletionProtection`.
    * Various data operation:
      * `Values` in the `Vector` type are now a pointer to allow flexibility when working with sparse-only indexes.
  </Warning>
</Update>

<Update label="2025-02-07" tags={["SDK"]}>
  ### Released .NET SDK v3.0.0

  Released [`v3.0.0`](https://github.com/pinecone-io/pinecone-dotnet-client/releases/tag/3.0.0) of the [Pinecone .NET SDK](/reference/sdks/dotnet/overview). This version uses the latest stable API version, `2025-01`, and adds support for [sparse-only indexes](/guides/index-data/indexing-overview#indexes-with-sparse-vectors).

  <Warning>
    Before upgrading to `v3.0.0`, update all relevant code to account for the following [breaking changes](/reference/api/versioning#breaking-changes). See the [`v3.0.0`](https://github.com/pinecone-io/pinecone-dotnet-client/releases/tag/3.0.0) release notes for full details.

    * [`embed` operation](/reference/api/2025-01/inference/generate-embeddings):
      * The `Embedding` type has changed from a simple object to a discriminated union, supporting both `DenseEmbedding` and `SparseEmbedding`. New helper methods available on the Embedding type: `IsDense` and `IsSparse` for type checking, `AsDense()` and `AsSparse()` for type conversion, and `Match()` and `Visit()` for pattern matching.
      * The `Parameters` property now uses `Dictionary<string, object?>?` instead of `EmbedRequestParameters`.

    * `rerank` operation:
      * The `Document` property now uses `Dictionary<string, object?>?` instead of `Dictionary<string, string>?`.
      * The `Parameters` property now uses `Dictionary<string, object?>?` instead of `Dictionary<string, string>?`.
  </Warning>
</Update>

## January 2025

<Update label="2025-01-29" tags={["General"]}>
  ### Update to the API keys page

  Added the **Created by** column on the [API keys page](https://app.pinecone.io/organizations/-/projects/-/keys) in the Pinecone Console. This column shows the email of the user who created the API key.
</Update>

<Update label="2025-01-29" tags={["Database"]}>
  ### Sparse-only indexes in early access

  You can now use [sparse-only indexes](/guides/index-data/indexing-overview#indexes-with-sparse-vectors) for the storage and retrieval of sparse vectors. This feature is in [early access](/release-notes/feature-availability).
</Update>

<Update label="2025-01-22" tags={["Assistant"]}>
  <img className="block max-w-full" noZoom src="https://mintcdn.com/pinecone/r0TaYXrfSrAYZYUj/images/release-notes/assistant-ga.png?fit=max&auto=format&n=r0TaYXrfSrAYZYUj&q=85&s=aa42277fa4b0cb3fafd1544c9b65ee24" width="1774" height="994" data-path="images/release-notes/assistant-ga.png" />

  Pinecone Assistant is generally available (GA) for all users.

  [Read more](https://www.pinecone.io/blog/pinecone-assistant-generally-available) about the release on our blog.
</Update>

<Update label="2025-01-09" tags={["SDK"]}>
  ### Released Node SDK v4.1.0

  Released [`v4.1.0`](https://github.com/pinecone-io/pinecone-ts-client/releases/tag/4.1.0) of the [Pinecone Node.js SDK](/reference/sdks/node/overview). This version adds support for [index tags](/guides/manage-data/manage-indexes#configure-index-tags) when creating or configuring indexes. It also adds a new `RetryOnServerFailure` class that automatically retries asynchronous operations with exponential backoff when the server responds with a `500` or `503` [error](/reference/api/errors).
</Update>

<Update label="2025-01-09" tags={["General"]}>
  ### New Billing Admin user role

  Added the Billing Admin [user role](/guides/organizations/understanding-organizations#organization-roles). Billing Admins have permissions to view billing details, usage details, and support plans.
</Update>

<Update label="2025-01-07" tags={["SDK"]}>
  ### Released Go SDK v2.2.0

  Released [`v2.2.0`](https://github.com/pinecone-io/go-pinecone/releases/tag/v2.2.0) of the [Pinecone Go SDK](/reference/sdks/go/overview). This version adds support for [index tags](/guides/manage-data/manage-indexes#configure-index-tags) when creating or configuring indexes.
</Update>