ceven

Studio By Ai21 Labs

Deploys and manages AI assistants, indexes live website data for grounding, and automates the creation of custom model routes to power complex text generation workflows.

Try Studio By Ai21 Labs in Ceven

Ask Ceven anything
Standard

Why use Ceven?

  1. AI native Studio By Ai21 Labs integration

    • Describe the outcome and Ceven picks the right Studio By Ai21 Labs 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 Studio By Ai21 Labs data, across all 57 of its actions.

  2. Managed auth

    • Built in OAuth with automatic token refresh and rotation.
    • One place to manage, scope, and revoke Studio By Ai21 Labs 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 Studio By Ai21 Labs, when, and on whose behalf.
    • The agent pauses and asks when Studio By Ai21 Labs is unclear instead of plowing ahead.

  4. Enterprise grade security

    • Fine grained access so you control which agents and people can reach Studio By Ai21 Labs.
    • Least privilege by default, read scopes first and only the writes a workflow needs.
    • A full audit trail of every Studio By Ai21 Labs action to support review and sign off.

Supported tools

Every action Ceven's agents can run on Studio By Ai21 Labs, and when to use it.

Create Assistant
Use this to set up a new AI assistant with custom configuration and behavior settings.
Create Website Connector
Use this to index website content via sitemap or specific URLs for grounding purposes.
Ingest Website Connector
Trigger the processing and indexing of website pages from a sitemap URL.
Create Assistant Route
Define a new conversational path or query pattern for an existing assistant.
Create Assistant Plan
Add a new plan containing Python code to guide assistant behavior.
Create Secret
Securely store sensitive information like API keys or passwords within the workspace.
Get Assistant
Pull details about a specific assistant including its tools and configuration.
Get Output Explanation
Retrieve the explanation for a specific execution to debug model reasoning.
List Assistants
Pull all available assistants filtered by workspace IDs.
Get Website Connector Status
Check the processing status of a website ingestion job.
Generate Requirements
Break down a high level task description into specific technical requirements.
Compare Text
Analyze changes between an original text string and a modified version.
Check Can Iframe
Tool to check if a URL can be embedded in an iframe. Use when you need to verify iframe compatibility before embedding external content.
Create Aftersales Parts Classification Batch
Tool to create a batch job for Fnac Aftersales parts classification. Use when you need to classify multiple aftersales cases to determine which parts are needed for repairs.
Create Demo
Tool to create a new demo in AI21 Studio. Use when you need to set up a new demo with specified visibility and status.
Generate Thread Name
Tool to generate a thread name from a query or conversation context. Use when you need to create a descriptive title for a conversation thread.
Create Kirsh Grant Compliance Preview
Tool to preview grant compliance for Kirsh grants. Use when you need to check compliance status for specific grant IDs.
Kirsh Grant Metadata Preview
Tool to retrieve grant metadata from the Kirsh preview endpoint. Use when you need to fetch grant information including grant name and associated organization details.
Create MCP Storage
Tool to create MCP (Model Context Protocol) storage in AI21 Studio. Use when you need to register and configure an MCP server connection for your workspace.
Check Kirsh Grant Compliance
Tool to check compliance for Kirsh grant IDs. Use when you need to verify compliance status for one or more grants.
Delete Assistant
Tool to delete an assistant by its unique ID. Use when you need to permanently remove an assistant from the system.
Delete Assistant Route
Tool to delete a route from an AI21 Studio assistant. Use when you need to remove a routing configuration from an assistant.
Delete Demo
Tool to delete a demo by its unique ID. Use when removing a specific demo from AI21 Studio.
Delete MCP Storage
Tool to delete an MCP (Model Context Protocol) storage configuration by its unique ID. Use when you need to permanently remove an MCP server configuration from AI21 Studio.
Delete Secret
Tool to delete a secret by its unique ID. Use when you need to permanently remove a secret from AI21 Studio.
Delete Website Connector
Permanently delete a website connector by its unique identifier. Use when you need to remove a website connector that is no longer needed.
Download Modified Document
Tool to download a modified document from AI21 Labs Studio. Use when you need to generate and download a document with specific content and filename.
Get Assistant Route
Tool to retrieve details of a specific route for an AI21 Studio assistant. Use when you need to get information about a particular route configuration.
Get Assistants By MCP
Tool to retrieve all assistants associated with a specific MCP (Model Context Protocol) storage configuration. Use when you need to list assistants that are connected to a particular MCP server.
Get Batch Prediction Status
Tool to check the status of a Fnac Aftersales parts classification batch prediction job. Use when you need to monitor the progress or retrieve results of a previously submitted batch.

30 actions · scroll to see them all

Frequently asked questions

Ceven manages the full pipeline by first creating a website connector and then triggering the ingestion process via sitemap or individual URLs. The agent does not just fire the request and stop; it polls the website connector status endpoint to ensure the pages are fully processed before marking the workflow as complete. This ensures that any AI assistant relying on that data for grounding has the most recent information. If the ingestion fails for specific URLs, Ceven identifies the failed links and can attempt a retry or notify the user of the specific pages that are blocked by robots txt files.
Yes. Ceven can use the Create Assistant Plan action to upload Python code that defines the operational logic of your AI21 Studio assistant. This allows the agent to move beyond simple prompting and into complex workflow orchestration where the model follows a coded plan to achieve a goal. You can provide the Python script as part of the prompt or have Ceven generate the script based on your requirements. Once the plan is created, the assistant follows those instructions during execution, making it possible to build highly structured tools for data analysis or complex customer interactions.
In AI21 Studio, an assistant is the primary entity that holds the configuration, tools, and general personality of the bot. A route is a more fine grained configuration that tells the assistant how to handle specific types of queries. Ceven uses routes to build decision trees; for example, one route might handle billing questions while another handles technical troubleshooting. By creating multiple routes, the agent ensures that the model uses the most relevant context and tools for the specific user intent, which significantly reduces hallucinations and improves the accuracy of the generated responses.
Ceven utilizes the Create Secret tool to ensure that sensitive credentials are never stored in plain text within your workflow logs or prompts. When a workflow requires an external API key to function, the agent pushes that key into the AI21 Studio secret store. The model then references the secret by name during execution rather than having the raw key passed in the request body. This architecture ensures that your credentials remain encrypted at rest and are only accessible to the model during the actual runtime of the assistant, following security best practices for enterprise AI deployments.
Ceven uses the Get Output Explanation tool to peek into the reasoning process of the model. When a response is flagged as incorrect, the agent retrieves the cached explanation for that specific execution. It then analyzes the steps the model took to arrive at the answer, comparing the output against the indexed website content. If the agent finds that the model ignored a specific piece of grounding data, it can automatically suggest a modification to the assistant plan or create a new route to better guide the model in the future, effectively creating a closed loop for quality improvement.
While Ceven can automate the creation of assistants, you are still bound by the quota limits of your AI21 Studio account tier. A common quirk is that the number of active assistants and the total number of indexed pages in website connectors are capped based on your billing plan. If you hit these limits, the API will return a quota error which Ceven will capture. The agent can then help you manage your workspace by listing all assistants and identifying old or unused ones to delete, freeing up capacity for new deployments without requiring you to manually audit your dashboard.
The AI21 Studio website connector works best with static HTML and sitemaps. If your website relies heavily on client side rendering via JavaScript, the connector may fail to see the content, as it primarily indexes the server side response. Ceven can help mitigate this by allowing you to ingest specific URLs individually if the sitemap is incomplete. However, users should be aware that content behind login walls or content that requires complex user interaction to load will not be indexed. The agent can verify the status of each URL to tell you exactly which pages were successfully ingested.
Yes, Ceven can use the Create Demo tool to spin up a shareable interface for any assistant you have configured. This is useful for rapid prototyping where you want to test a specific route or plan before moving it to production. The agent can configure the visibility settings to ensure the demo is only accessible to your team or make it public for client reviews. By combining the Create Assistant and Create Demo actions, Ceven can take a raw prompt and turn it into a functional, hosted AI application in a single automated sequence.

Alternatives to Studio By Ai21 Labs

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

OpenAI logoOpenAIAnthropic logoAnthropicCohere logoCohere

Try Ceven on your stack

Plug Ceven on top of the tools you already run. Connect Studio By Ai21 Labs 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