Data Source

Pinecone

Semantic search and vector storage for AI workflows

Pinecone is the leading vector database for AI applications. Store high-dimensional embeddings alongside metadata, then search them by semantic similarity in milliseconds. CipherSense Agents connects directly to the Pinecone REST API using your API key — no proxy required. Use Pinecone nodes to build RAG (Retrieval-Augmented Generation) pipelines, semantic search, recommendation engines, and long-term AI memory. Note: Pinecone also offers an official MCP server (@pinecone-database/mcp) for local use. If you prefer the MCP approach, connect it via the Custom MCP integration instead.

Official documentation

Setup

Get your Pinecone API key

Sign in to the Pinecone console and go to API Keys. Click 'Create API key' or copy an existing one. The key starts with pcsk_.

Pinecone Console

Find your Index Host URL

In the Pinecone console, open the index you want to use. The Host URL is shown in the index details panel — it looks like https://my-index-abc123.svc.pinecone.io. Copy it exactly. You need this for all data-plane operations (query, upsert, fetch, delete).

Add the integration in CipherSense Agents

Go to Project > Integrations > Add Integration > Pinecone. Paste your API key and index host URL. Optionally set a default namespace. Click Save & Test to verify the connection.

Use in a workflow

Drag a Pinecone node into the Visual Designer. For RAG: use an LLM node to generate an embedding → feed it to a Pinecone query node → pass the top matches as context to the next LLM node. For ingestion: generate embeddings → upsert to Pinecone with metadata.

Connection Fields

Fields required when adding this integration in your Project › Integrations.

Field	Required	Description
API Key	Required	Your Pinecone API key from the Pinecone console > API Keys.
Index Host URL	Optional	Full host URL of your target index. Required for query, upsert, fetch, delete, and describe_index_stats operations. Can also be set per-node.
Default Namespace (optional)	Optional	Default namespace for vector operations. Namespaces allow you to partition vectors within one index. Can be overridden per node.

Common Use Cases

RAG pipeline: chunk document → generate embeddings (OpenAI/Claude) → upsert to Pinecone → query at runtime → inject top-k results as LLM contextSemantic search: user query → embed → Pinecone query → return matched document metadataLong-term AI memory: after each conversation → embed summary → upsert with session metadata → recall relevant memories at start of next sessionRecommendation engine: embed user profile → query Pinecone → return most similar itemsDuplicate detection: embed incoming content → query with high topK → filter matches above thresholdKnowledge base Q&A: ingest docs → query on user question → pass top matches to LLM for grounded answers

Ready to connect Pinecone?

Add this integration from your project dashboard and wire it into a workflow.

Open Dashboard ← Back to Integrations