ceven

Needle

Connects your proprietary data collections to AI workflows, enabling semantic search and retrieval augmented generation for your agents.

Try Needle in Ceven

Ask Ceven anything
Standard

Why use Ceven?

  1. AI native Needle integration

    • Describe the outcome and Ceven picks the right Needle calls, fills the parameters, and checks the result.
    • Structured, agent friendly tool schemas so each call runs reliably instead of by guesswork.
    • Rich coverage for reading, writing, and querying your Needle data, across all 16 of its actions.

  2. Managed auth

    • Built in OAuth with automatic token refresh and rotation.
    • One place to manage, scope, and revoke Needle access.
    • Per user and per environment credentials instead of shared keys.

  3. Agent optimized design

    • Actions are tuned from real success and error rates so reliability climbs over time.
    • Full execution logs so you always know what ran in Needle, when, and on whose behalf.
    • The agent pauses and asks when Needle is unclear instead of plowing ahead.

  4. Enterprise grade security

    • Fine grained access so you control which agents and people can reach Needle.
    • Least privilege by default, read scopes first and only the writes a workflow needs.
    • A full audit trail of every Needle action to support review and sign off.

Supported tools

Every action Ceven's agents can run on Needle, and when to use it.

Create Collection
Use this when you need to start a new knowledge base for a specific project or client after confirming the collection name.
Delete files from collection
Remove specific documents by file id. Use this to purge outdated information or incorrect data from a collection.
Get collection stats
Pull document counts, index size, and timestamps for a specific collection id to monitor growth and freshness.
List Collection Files
Retrieve a paginated list of all files within a specific collection id to audit stored content.
List Collections
Pull a list of all available collections to find the correct id for a specific knowledge domain.
Semantic Search
Query a collection using natural language to find the most relevant document chunks based on meaning.
Upload Document
Push a new file into a collection for indexing and future retrieval in RAG workflows.
Update Collection Metadata
Modify the tags or descriptions associated with a collection to improve organization.
Get File Content
Pull the raw text or metadata for a specific file id within a collection.
Delete Collection
Permanently remove an entire collection and all its indexed files when the project ends.
Check Index Status
Verify if a recently uploaded document has finished processing and is ready for search.
Clear Collection
Remove all files from a collection while keeping the collection shell and settings intact.
Delete files from a collection
Tool to delete files from a specific collection by providing file ids. use after confirming valid file ids to remove from the collection.

13 actions · scroll to see them all

Frequently asked questions

Ceven acts as a secure bridge between your data and the Needle API. We do not store your indexed documents on our own servers. When a workflow triggers a search, Ceven sends the query to Needle, retrieves the relevant chunks, and passes them to the LLM. All communication happens over encrypted channels. You maintain full ownership of the collections in your Needle account. If you delete a collection or a file in Needle, it is immediately unavailable to the Ceven agent. We follow the principle of least privilege, only requesting the scopes necessary to perform the specific search or management tasks defined in your workflow.
Yes. You can build a Ceven workflow that monitors a folder in Google Drive or a GitHub repository. Whenever a new document is added or an existing one is modified, the agent can trigger the Upload Document action in Needle. To avoid duplicates, the agent can first use List Collection Files to check for existing versions. This ensures your AI agent always has the most current information without manual intervention. You can even set up a weekly audit workflow that lists all files and flags those that have not been updated in over ninety days for human review.
Ceven allows you to configure the top k parameter for every Needle search action. This limits the number of document chunks returned to the LLM. If the search returns too much noise, you can refine the workflow to use more specific metadata filters or implement a reranking step. The agent can also be instructed to perform multiple smaller searches across different collections and synthesize the results. This prevents the LLM context window from being overwhelmed and reduces the chance of the model ignoring the most important piece of information located in the middle of a large text block.
Needle has different tier gating based on your subscription plan. The free tier has a strict limit on the total number of tokens and documents you can index across all collections. If you hit these limits, the Create Collection and Upload Document actions will return an error. You can use the Get collection stats action in Ceven to monitor your usage in real time and set up an alert workflow that notifies you via Slack or email when you reach eighty percent of your capacity. Upgrading your Needle plan immediately increases these limits without requiring any changes to your Ceven workflows.
Keyword search looks for exact word matches, while Needle semantic search uses vector embeddings to understand the intent and meaning behind a query. For example, if you search for 'how to fix power issues', Needle will find documents discussing 'voltage drops' or 'electrical failures' even if the word 'power' never appears. Ceven leverages this by allowing you to build agents that understand user intent. You can set up a workflow that tries a semantic search first and falls back to a keyword search if the confidence score is too low, ensuring the most accurate retrieval possible.
Ceven supports multiple connections for Needle. You can connect different accounts for different departments or clients. In your workflow, you simply specify which connection to use for a particular action. This is particularly useful for agencies managing RAG implementations for multiple customers. You can create a master workflow that routes queries to different Needle accounts based on the customer email address in the incoming ticket. Each connection is isolated, meaning an agent using the Account A connection can never accidentally access or modify collections residing in Account B.
The total latency depends on the complexity of the query and the size of the Needle collection. Typically, a semantic search in Needle takes a few hundred milliseconds. The bulk of the time is spent by the LLM generating the final response based on the retrieved text. To optimize this, Ceven supports asynchronous processing for background tasks like indexing. For user facing bots, we recommend using smaller, focused collections in Needle rather than one giant monolithic index, as this improves retrieval speed and increases the precision of the context provided to the AI model.
Needle is designed for indexing and retrieval rather than real time stream processing. While you can upload documents frequently, there is a short processing delay while Needle generates the embeddings for the new text. This means a document uploaded one second ago might not appear in search results for another few seconds. If your workflow requires absolute real time consistency, we recommend implementing a check using the Check Index Status action to ensure the document is fully processed before the agent attempts to retrieve it for a customer response.

Alternatives to Needle

Other tools that solve a similar problem. Ceven supports these too, so you can switch or run more than one at once.

Pinecone logoPineconeWeaviate logoWeaviateMilvus logoMilvus

Try Ceven on your stack

Plug Ceven on top of the tools you already run. Connect Needle and the rest of your stack, describe the outcome, and its agents handle the work end to end, days of it in minutes.

Get started for free