ceven

OpenAI

Connects your proprietary data to OpenAI assistants, automates the creation of conversation threads, and manages the lifecycle of fine tuning jobs and file uploads.

Try OpenAI in Ceven

Ask Ceven anything
Standard

Why use Ceven?

  1. AI native OpenAI integration

    • Describe the outcome and Ceven picks the right OpenAI 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 OpenAI data, across all 126 of its actions.

  2. Managed auth

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

  4. Enterprise grade security

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

Supported tools

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

Create Assistant
Use this when you need to build a new AI persona with specific instructions, a chosen model, and a set of enabled tools.
Create Message
Add a user or system message to an existing conversation thread. Use this to feed data into an active chat session.
Create Thread
Initialize a new conversation session. Use this to start a fresh interaction before adding messages or running the assistant.
Delete Assistant
Remove an assistant permanently by ID. Use this to clean up old versions or retired personas.
Delete File
Remove an uploaded file from the OpenAI platform. Use this to maintain data privacy or clear out old training sets.
List Files
Pull a list of all files currently uploaded to the organization. Use this to verify if a document exists before referencing it.
List Fine Tunes
Retrieve all fine tuning jobs. Use this to track the progress of model training runs for specialized tasks.
List Models
Pull all available models your account can access. Use this to check for new model releases or availability.
List Run Steps
Inspect the individual steps a model took to reach a conclusion. Use this for debugging complex assistant logic.
Modify Thread
Update the metadata of an existing thread. Use this to tag conversations with customer IDs or project codes.
Retrieve Assistant
Pull the current configuration and instructions of a specific assistant to verify its setup.
Retrieve Model
Fetch detailed metadata for a specific model ID to confirm its capabilities before deployment.
Retrieve Thread
Pull the current state and metadata of a conversation thread by its unique ID.
Upload File
Upload a document for use with assistants or fine tuning. Use this to provide the model with custom knowledge.

14 actions · scroll to see them all

Frequently asked questions

Ceven allows you to provide your own OpenAI API key which is stored using AES 256 encryption at rest. We never share this key with other users or use it for any purpose other than executing the specific actions you define in your workflows. The key is injected into the request header only at the moment of the API call and is not logged in our system. You can rotate your key at any time in the OpenAI dashboard and update it within the Ceven connection settings to ensure your environment remains secure. This gives you full control over your billing and usage limits while allowing our agents to operate on your behalf.
Yes. Ceven can trigger the creation of fine tuning jobs by uploading the required training files and initiating the run. The agent can then periodically call the list fine tunes endpoint to monitor the status of the job. Once the job is marked as succeeded, Ceven can update your assistant configuration to use the newly created fine tuned model. This creates a closed loop where the agent identifies a need for better performance, triggers the training, and deploys the updated model without manual intervention. You still maintain full visibility into the costs and logs within your OpenAI account.
Ceven follows the standard OpenAI API limits for file uploads. Currently, OpenAI imposes a limit on the total storage per organization and individual file size limits. One specific quirk is that files uploaded for the Assistants API are stored differently than those for fine tuning. If you upload a file for a knowledge base, it must be formatted correctly for the retrieval tool to function. Ceven helps by validating the file extension before upload, but if you hit a rate limit during a bulk upload, the agent will automatically implement an exponential backoff strategy to ensure all documents are eventually processed.
OpenAI threads manage the conversation history automatically, but they are still subject to the context window of the underlying model. Ceven manages this by allowing you to retrieve and prune threads or start new ones when a conversation reaches a certain length. The agent can summarize the previous thread and pass that summary as the first message in a new thread to maintain continuity. This prevents the model from losing the original goal of the conversation while avoiding the high token costs and potential latency associated with extremely long threads that push the limits of the model context.
Yes. Because Ceven can list and retrieve model metadata, you can build logic that routes requests to different models based on the task. For example, you can use GPT 4o for complex reasoning tasks and GPT 4o mini for simple classification to save on costs. The agent checks the availability of the model in your account before making the call. If a specific model is unavailable or you have reached a tier limit for that model, the workflow can be configured to fail over to a different model version to ensure your business process does not stop.
The Chat Completions API is stateless, meaning you must send the entire history back to OpenAI with every request. The Assistants API is stateful, using threads to store history on the OpenAI servers. Ceven leverages the Assistants API for complex, long running workflows because it reduces the amount of data the agent needs to pass back and forth. This makes it easier to build tools that remember user preferences across multiple days. Ceven manages the thread IDs and assistant IDs in the background so you can focus on the logic of the conversation rather than the plumbing of the API.
Yes. Ceven can interact with OpenAI models that support vision by passing image URLs or base64 encoded images within the message payload. You can build workflows that analyze screenshots, read receipts, or inspect product photos and then trigger downstream actions based on the visual analysis. The agent can combine this with the retrieval tool to compare an uploaded image against a set of technical documents. This allows for sophisticated quality control or automated auditing workflows where the AI acts as a visual inspector and a technical expert simultaneously.
OpenAI enforces strict rate limits based on your account tier. When Ceven encounters a 429 rate limit error, it does not simply fail the workflow. Instead, it uses a smart retry mechanism with jitter to spread out requests. If the limit is hit consistently, the agent will notify the workflow owner via your chosen notification channel and pause the queue. This prevents your API key from being temporarily blocked for abuse. We recommend upgrading your OpenAI tier if you notice frequent rate limit hits during high volume batch processing of files or messages.

Alternatives to OpenAI

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

Anthropic logoAnthropicGoogle Cloud Vertex AI logoGoogle Cloud Vertex AIMistral AI logoMistral AI

Try Ceven on your stack

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