ceven

Nanonets

Extracts structured data from unstructured documents and pushes the results into your downstream databases or ERP systems as they are processed.

Try Nanonets in Ceven

Ask Ceven anything
Standard

Why use Ceven?

  1. AI native Nanonets integration

    • Describe the outcome and Ceven picks the right Nanonets 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 Nanonets data, across all 11 of its actions.

  2. Managed auth

    • Built in OAuth with automatic token refresh and rotation.
    • One place to manage, scope, and revoke Nanonets 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 Nanonets, when, and on whose behalf.
    • The agent pauses and asks when Nanonets is unclear instead of plowing ahead.

  4. Enterprise grade security

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

Supported tools

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

Create OCR Model
Use this when you need to initialize a new ocr model before you start training it on your specific document types.
Delete OCR Model
Permanently remove a trained model by its id. Use this for cleaning up old versions or unused test models.
Get all OCR models
Pull a paginated list of all ocr models in the account. Use this to inspect available models after authentication.
Get All Prediction Files
Fetch all prediction files for a specific model. Use this to list all inference requests after processing is complete.
Get OCR Model Details
Pull the full metadata of a specific model by its id to check configuration and status.
Get OCR Training Images
Retrieve training images for a model. Use this to page through images before starting analysis or training.
Get Workflows
Retrieve a list of all workflows in the nanonets account to inventory configured automation paths.
List Workflow Documents
Pull a paginated list of documents processed by a specific workflow to view results after processing.
Update OCR Model
Modify a model name, categories, notes, or classification settings after reviewing the current configuration.
Upload Training Images by File
Upload training images as files to a specified ocr model to improve extraction accuracy.
Upload Training Images by URL
Add images to a model for training purposes by providing the direct urls of the files.

11 actions · scroll to see them all

Frequently asked questions

Ceven monitors the confidence score returned for every field extracted by Nanonets. You can set a specific threshold, such as eighty percent, in your workflow settings. If Nanonets returns a value below this number, the agent pauses the automatic sync to your database and instead creates a task in your project management tool or sends a Slack notification. This allows a human to review the document in the Nanonets dashboard and correct the field. Once the human marks the record as verified, Ceven detects the status change and completes the data push to your downstream system, ensuring high data integrity.
Ceven respects the rate limits imposed by your specific Nanonets plan. It is important to note that Nanonets enforces strict limits on the number of concurrent API requests for document uploads and model training. If you attempt to push thousands of documents at once, Nanonets may return a 429 too many requests error. To handle this, Ceven implements an exponential backoff strategy, meaning the agent will automatically pause and retry the request after a short delay. For very high volume accounts, we recommend batching uploads to avoid hitting these limits and causing delays in your real time workflows.
Yes. Ceven can automate the data collection phase of model training. You can build a workflow that pulls sample documents from an S3 bucket or Google Drive and uses the Nanonets upload training images action to feed them into a specific model. While the actual labeling of the data must be done by a human within the Nanonets interface to ensure accuracy, Ceven handles the heavy lifting of moving the files. Once you have labeled enough samples and triggered the training process in Nanonets, Ceven can monitor the model status and notify you when the new version is ready for production.
Ceven does not store your raw documents. The agent acts as a secure pipe that moves the file from your source, such as an email or cloud storage, to the Nanonets API. Once Nanonets processes the file and returns the structured JSON data, Ceven processes that data and passes it to your destination system. We only store the metadata and the transaction logs necessary to track the success or failure of the workflow. This architecture ensures that your sensitive documents remain within your controlled environments and the secure vault provided by Nanonets, minimizing your data footprint.
The integration is most effective for documents with a consistent structure, such as invoices, receipts, purchase orders, and identity documents. Because Nanonets uses a combination of OCR and deep learning, it handles variations in layout better than traditional template based tools. However, extremely blurry images or handwritten notes with poor legibility may result in lower confidence scores. We recommend using the Nanonets training actions via Ceven to upload at least ten to twenty examples of your most common document variants, which allows the AI to learn the specific nuances of your business documents before you go live.
Nanonets processes multi page documents as a single unit, and Ceven is configured to handle the resulting array of data. If a document has multiple tables across different pages, Nanonets extracts them as a continuous set of records. Ceven can then iterate through these records and create individual entries in your destination system. For example, if an invoice has three pages of line items, the agent will loop through every item extracted by Nanonets and create a separate row in your accounting software, ensuring that no data is lost regardless of the document length.
Absolutely. Ceven can interact with as many models as your Nanonets account supports. You can build a router workflow that inspects a document based on its filename or source and directs it to the correct Nanonets model. For instance, a file containing the word invoice in the name goes to the invoice model, while a file from the HR folder goes to the employee onboarding model. This allows you to centralize all your document processing through one Ceven agent while leveraging the specialized accuracy of multiple fine tuned Nanonets models for different business functions.
Yes, you can use the update OCR model action. This allows the agent to change the name of the model or update its classification categories via API. This is particularly useful if you are dynamically creating models for different clients or projects and need to keep the naming conventions in sync with your CRM. While the deep training and labeling must happen in the Nanonets UI, the administrative management of the model properties can be fully automated through Ceven, reducing the time you spend switching between different browser tabs.

Alternatives to Nanonets

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

Rossum logoRossumABBYY logoABBYYHyperscience logoHyperscience

Try Ceven on your stack

Plug Ceven on top of the tools you already run. Connect Nanonets 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