# OpenAI MCP for AI Agents AI Agent Connect > OpenAI MCP. Manage models, fine-tunes, batches, and assistants through your AI agent. Stop jumping between tabs to check job statuses or audit files. You can list models, track batch progress, and manage files for fine-tuning directly in your IDE or chat interface. ## Overview - **Category:** ai-frontier - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_nUi0FAK4SIEESzN3EN6rUcUtM1G2aBwB1LndkjlH/ai-agent-connect - **Tags:** llm-management, fine-tuning, model-discovery, ai-assistants, file-management, api-orchestration ## Description You're tired of keeping twenty browser tabs open just to see if your latest fine-tuning job actually succeeded or if your batch processing is still running. This OpenAI MCP changes that by pulling your entire resource management suite into your existing AI workflow. Instead of hunting through the dashboard, you can just ask your agent to give you a status update or tell you which files are currently taking up space in your account. It turns your AI into a hands-on ML ops assistant that handles the grunt work of resource auditing and job tracking. Because Vinkius hosts this and thousands of other tools, you can keep your entire infrastructure under one roof without having to manage separate connections for every single service. You get to stay in your flow state while your agent handles the tedious task of checking permissions, listing available models, and cleaning up old training data. This means you can audit your account, monitor long-running processes, and manage your model assets without ever leaving your IDE or chat interface. It removes the friction of manual monitoring and lets you spend your time on actual model development. Whether you are scaling up production workloads or just trying to organize your workspace, this connection provides a direct line to your OpenAI account. It's about getting the information you need exactly when you need it, rather than going on a scavenger hunt through various management consoles. ## Tools ### cancel_batch Stop a running batch job immediately by providing the batch ID. This is useful for stopping jobs that are no longer needed. ### cancel_fine_tune Stop a long-running or incorrect fine-tuning job to save on costs. The job status will change to cancelled. ### create_batch Start a new batch processing job with a specific input file and endpoint. It returns the batch ID for tracking. ### delete_file Remove an uploaded file from your account to free up space. This action is irreversible. ### get_batch Get the current status and details of a specific batch job. Use this to check progress on a per-job basis. ### get_fine_tune Retrieve the specific status and metadata for a fine-tuning job. Useful for checking training completion. ### get_model Check the metadata, owner, and permissions for a specific model. Use this to verify a model exists before use. ### list_assistants View all configured Assistants including their instructions and tools. Use this to audit your agent fleet. ### list_batches See all active and completed batch jobs in your account. Each shows its status and request counts. ### list_files Browse all uploaded files used for fine-tuning or the Assistants API. You can filter these by purpose. ### list_fine_tunes View your entire fine-tuning pipeline, including estimated finish times. Use this to monitor your training progress. ### list_models See every model available to your account along with their capabilities. Use this to discover available options. ### get_assistant Pull specific details for a single OpenAI Assistant using its ID. This shows model info and instructions. ## Prompt Examples **Prompt:** ``` What's the status of my training jobs? ``` **Response:** ``` I've checked your fine-tuning jobs. You have one active job (ftjob-9921) which is currently 'running' and about 45% complete. It's expected to finish in about 40 minutes. You also have two completed jobs from last week. ``` **Prompt:** ``` List my uploaded files for the assistants. ``` **Response:** ``` You have 4 files currently uploaded for your Assistants: | Filename | Purpose | Size | | :--- | :--- | :--- | | handbook.pdf | Assistants | 12MB | | policy_v2.pdf | Assistants | 8MB | | user_data.jsonl | Fine-tune | 2MB | | training_v1.jsonl | Fine-tune | 15MB | ``` **Prompt:** ``` Show me all the models I can use. ``` **Response:** ``` Here are the models available in your account: * **gpt-4o**: Our flagship high-intelligence model. * **gpt-4o-mini**: Great for fast, cost-effective tasks. * **whisper-1**: Best for speech-to-text. * **text-embedding-3-small**: Optimized for search and similarity. * **dall-e-3**: For high-quality image generation. ``` ## Capabilities ### Check fine-tuning status See if your training jobs are running, succeeded, or failed in real time. ### Manage batch jobs Create and track large-scale API requests without manual monitoring. ### Audit files Browse and delete uploaded files for Assistants or fine-tuning. ### Inspect Assistants View configurations, models, and instructions for your custom agents. ### Discover models List every model available to your organization and check their metadata. ## Use Cases ### Stopping a failed training run An ML engineer realizes they uploaded a corrupted file. They ask their agent to use cancel_fine_tune to stop the job immediately. ### Starting a high-volume batch job A dev needs to process 10,000 requests cheaply. They ask their agent to use create_batch to start the job and monitor the progress. ### Auditing Assistant instructions A product lead wants to check what tools their bots have. They ask their agent to use list_assistants to review the fleet. ### Getting specific bot details A developer needs to check a specific bot's model. They ask their agent to use get_assistant to pull the configuration. ## Benefits - Get real-time training updates by using list_fine_tunes to see your progress without opening a browser. - Stop wasting money on incorrect runs by using cancel_batch to halt jobs that aren't performing as expected. - Clean up your storage easily by using list_files to see your uploads and delete_file to remove old data. - Audit your agent fleet in seconds by using list_assistants to review configurations across your entire account. - Discover new capabilities instantly by using list_models to see every option available to your organization. ## How It Works The bottom line is you get a unified command center for your OpenAI resources inside your favorite AI tool. 1. Subscribe to the OpenAI MCP on the Vinkius Marketplace. 2. Add your OpenAI API Key to your AI client settings. 3. Ask your agent to list your fine-tuning jobs or check batch statuses. ## Frequently Asked Questions **Can I use the OpenAI MCP to manage my fine-tuning jobs?** Yes, you can use the OpenAI MCP to monitor, list, and cancel fine-tuning jobs. It lets you see the status of your training pipeline directly through your AI client. **How does the OpenAI MCP help with batch processing?** It allows you to create, track, and cancel batch jobs using natural language. You can check the status of large-scale API requests without leaving your workspace. **Can I delete files from my OpenAI account using this Connector?** Yes, you can use the delete_file tool via the OpenAI MCP to remove uploaded files. This helps you manage your storage and keep your account clean. **Does the OpenAI MCP work with Cursor or Claude?** Yes, the OpenAI MCP works with any MCP-compatible client, including Claude, Cursor, and Windsurf. You just connect it via Vinkius and start managing your resources. **How do I see which models are available in my OpenAI account?** You can ask your agent to list all available models. The OpenAI MCP will pull the latest data on model IDs, owners, and permissions for you. **Can I use the OpenAI MCP to audit my Assistants?** Yes, it provides a way to list all your Assistants and inspect their specific instructions, models, and tools in one view. **How do I get my OpenAI API Key?** Log in to the [**OpenAI Platform**](https://platform.openai.com/api-keys), go to **API Keys** in the left sidebar, click **Create new secret key**, give it a name and copy the key immediately — it starts with `sk-proj-` and won't be shown again. **Can I monitor my fine-tuning jobs?** Yes! Use `list_fine_tunes` to see all fine-tuning jobs with their status (validating_files, queued, running, succeeded, failed, cancelled). Use `get_fine_tune` with a specific job ID for detailed info including training progress, estimated finish time and result model ID. You can also cancel running jobs with `cancel_fine_tune`. **Can I manage batch processing jobs?** Yes! Use `list_batches` to see all batch jobs, `create_batch` to submit new batches with an input file ID and endpoint, `get_batch` to check progress and `cancel_batch` to stop running jobs. Batches process requests asynchronously at a lower cost than individual API calls. **Can I list and inspect my Assistants?** Yes! Use `list_assistants` to see all configured Assistants with their models, tools (code interpreter, file search, function calling) and instructions. Use `get_assistant` with a specific assistant ID for full details including file IDs and metadata.