What is Pinecone?
Quick answer
Pinecone is a hosted vector database: a service that stores embeddings, the numeric fingerprints of your text, and returns the pieces closest in meaning to a question. Automation tools like n8n use it so an AI agent can search a handbook or a pile of documents before it answers.
Last updated
Updated · By Robert Breen
Why it matters for a small business
A chatbot that answers from your own documents needs somewhere to keep them in a searchable form. Pinecone is one of the most common choices because there is nothing to install: you create an index in the web console, copy an API key, and tools such as n8n read and write to it. That makes it a practical first vector database for an office team building its first RAG chatbot.
As of October 2026, the free Starter plan is enough to learn on: up to 5 indexes, with Serverless indexes on AWS in a handful of regions such as Virginia (us-east-1). Paid plans add more indexes, other clouds and higher limits. The thing to get right is the index setup: its Dimension must match the embeddings model you send, and you can't change it after the index is created.
In a real lesson: Make Your Handbook Searchable for an AI Helpdesk using Pinecone + n8n
In the Pinecone + n8n load lesson you run operations at Cedar Ridge Partners, a made-up consulting firm. In the Pinecone console you click Create index, name it stepthrough-tests and pick text-embedding-3-small under Configuration. Pinecone fills in Metric: cosine, but Dimension shows 512, while n8n's Embeddings OpenAI node sends 1,536 numbers per chunk. You change it to 1536, leave Serverless, AWS and Virginia (us-east-1), and create the index.
An n8n workflow then loads the employee handbook PDF into it through the Pinecone Vector Store node in Insert Documents mode, and the index browser shows 38 records, one per chunk.
In the chatbot lesson the same node becomes a tool for an AI Agent, in Retrieve Documents (As Tool for AI Agent) mode with Limit 4, so every answer starts with a Pinecone search.
Try this lesson free or read the step-by-step guide.
Common confusions
Pinecone vs a vector database
A vector database is the type of tool. Pinecone is one product of that type, hosted for you. Others include Qdrant, Weaviate and Supabase with pgvector, and n8n has nodes for several of them.
Pinecone vs Pinecone Assistant
The Pinecone console also offers an Assistant that takes uploaded files and answers questions itself. The lessons use the plain index, so the n8n agent, its rules and its citations stay under your control.
Tips
- Write down which embeddings model an index was built for, and use that same model to load and to search.
- Use a separate index or namespace per document set, so a search never mixes the handbook with sales notes.
- When you reload changed documents, clear the old records first or you'll store everything twice.
Related terms
More n8n and automation terms
Where you use it: free lessons
- Make Your Handbook Searchable for an AI Helpdesk using Pinecone + n8n (n8n, 14 min)
- Answer Staff Questions from Your Own Docs using n8n + Pinecone (n8n, 13 min)
More free tutorials
Frequently asked questions
- Is Pinecone free?
- It has a free Starter plan, which as of October 2026 allows up to 5 indexes, Serverless on AWS, with monthly read, write and storage limits. Creating a sixth index asks you to upgrade to a paid plan.
- What dimension should my Pinecone index use?
- The number your embeddings model produces. OpenAI's text-embedding-3-small, n8n's default, produces 1,536, so choose 1536. If the numbers don't match, inserts fail.
- Do I need to code to use Pinecone?
- No. You create the index in the web console, and n8n's Pinecone Vector Store node handles inserting and searching with your API key saved as a credential.