canonical: https://jentic.com/apis/bentoml.com/bentoml

# Bentoml BentoCloud API

The BentoCloud API deploys and operates machine-learning models in production: managing bentos and model repositories, creating deployments and endpoints on clusters, and scaling inference services. It covers organizations, clusters, secrets, and usage metrics for running models at scale. Requests authenticate with an API token sent in the X-YATAI-API-TOKEN header.

## For AI agents

Deploy and manage machine-learning inference endpoints on BentoCloud, list bentos and models, and scale deployments across clusters.

## Scope

Does not handle model training, data labeling, or foundation-model inference itself - use for deploying and operating model serving on BentoCloud only.

## Capabilities

- Deploy machine-learning models as inference endpoints on a cluster
- List and manage bentos and model repositories
- Create, start, and terminate serving endpoints
- Manage clusters, deployments, and their revisions
- Handle organization secrets, members, and usage metrics

## Use cases

### Model deployment automation

An MLOps pipeline needs to promote a packaged model to a serving endpoint. The BentoCloud API creates a deployment on a cluster from a bento and exposes it as an endpoint, so a pipeline can ship a new model version programmatically.

Example prompt: Create a deployment via POST `/api/v1/clusters/{clusterName}/deployments` from a selected bento

### Inference endpoint lifecycle

A platform team needs to start, scale, and retire serving endpoints as demand shifts. The endpoint operations create an endpoint, start it, and terminate it so capacity tracks load without manual console work.

Example prompt: List endpoints via GET `/api/v1/endpoints`, then terminate an idle one via POST `/api/v1/endpoints/{endpointUID}/terminate`

### Model and bento inventory

A registry integration needs to know which packaged models exist. The bento and model operations list all bentos and models across repositories so a backend can audit what is available to deploy.

Example prompt: Call GET `/api/v1/bentos` and GET `/api/v1/models` to inventory available artifacts

### AI agent model operations

An AI agent managing inference infrastructure can deploy and scale models without hand-wiring the BentoCloud token. Through Jentic the agent discovers the deployment operations by intent and calls them with the stored credential.

Example prompt: Search Jentic for 'deploy a model endpoint', load the schema, and create a deployment on a cluster

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| GET | `/api/v1/deployments` | List deployments in the organization |
| POST | `/api/v1/clusters/{clusterName}/deployments` | Create a deployment on a cluster |
| GET | `/api/v1/bentos` | List all bentos |
| GET | `/api/v1/models` | List all models |
| GET | `/api/v1/endpoints` | List organization endpoints |
| POST | `/api/v1/endpoints` | Create a serving endpoint |

## Key resources

- **Deployments** — Create, update, and terminate model deployments on clusters
- **Endpoints** — Create, start, and terminate serving endpoints
- **Bentos** — List and manage packaged model artifacts and repositories
- **Models** — List and manage model repositories and versions
- **Clusters** — Manage clusters and their members and resources

## AI readiness

This API is usable in Jentic One now. Its AI-readiness score against Jentic's framework shows where it stands today and where improvements would make it even easier for agents to use.

- **Score:** 30 / 100
- **Maturity:** Non-Ready
- **Dimensions:**
  - Foundational Compliance: 44 / 100
  - Developer Experience & Jentic Compatibility: 56 / 100
  - AI-Readiness & Agent Experience: 15 / 100
  - Agent Usability: 42 / 100
  - Security: 50 / 100
  - AI Discoverability: 31 / 100
- **View full report:** https://jentic.com/apis/bentoml.com/bentoml/scorecard
- **How the score is calculated:** https://docs.jentic.com/reference/api-readiness-framework/overview/
- **More about the dimensions:** https://docs.jentic.com/reference/api-readiness-framework/specification/#dimensional-model-overview

### Score it yourself

Every API in the directory is allowlisted, so you can re-score it with no key required.

- **Score your own API:** https://jentic.com/scorecard.md
- **Scoring CLI agent skill:** https://github.com/jentic/jentic-api-scorecard/blob/main/skills/jentic-api-scorecard/SKILL.md

```sh
npx @jentic/api-scorecard-cli score <openapi-url>
```

## Why Jentic

- **Setup:** Wiring the BentoCloud API by hand means setting the X-YATAI-API-TOKEN header on every request and shaping the deployment, endpoint, and cluster payloads yourself. Through Jentic you install once, import the API from the API Directory, store your credential once, and your agent calls it.
- **Permission scoping:** The BentoCloud API addresses clusters, deployments, and endpoints by id in the URL path, so you can limit the agent to the operations it needs, such as listing bentos or creating a deployment. You credit the agent only with the operations you allow, so terminating a deployment stays out unless you add it.
- **Credential handling:** Your the BentoCloud credential is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'deploy a model endpoint', and Jentic returns the matching BentoCloud operation with its input schema so the agent calls the right endpoint without browsing the reference docs.

## Related APIs

- **Replicate API** — Run and host models without managing clusters yourself
- **Hugging Face API** — Hosted inference and a large model hub
- **OpenAI API** — Foundation-model inference for capabilities you do not self-host

## FAQ

### What authentication does the BentoCloud API use?

The BentoCloud API authenticates with an API token sent in the X-YATAI-API-TOKEN header. Through Jentic the token is stored encrypted in your own instance and injected into each request at run time.

### Is there a the BentoCloud API MCP server?

You don't need an MCP server to give your agent the BentoCloud API. Jentic connects it directly from the API Directory: import it, store your credential once, and your agent calls it. That keeps another server's tool definitions out of your agent's context while still giving it the operation set from the spec.

### Can the BentoCloud API deploy and scale model endpoints?

Yes. It creates deployments on clusters from packaged bentos, exposes them as endpoints, and lets you start and terminate endpoints so serving capacity can track demand.

### Can I limit what my agent is allowed to do with the BentoCloud API?

Yes. Because you run Jentic One yourself, your own rules decide which the BentoCloud API operations and credentials the agent may use. You grant only the operations it needs, such as deploying a model, so deleting a cluster stays off limits unless you add it. The choice of what the agent may call is yours.

### How do I connect my agent to the BentoCloud API through Jentic?

Search Jentic by intent such as 'deploy a model endpoint' to discover the operation, load its input schema, and your agent calls it with your stored credential injected at run time. To run it on your own infrastructure, install Jentic One from its GitHub repo.