canonical: https://jentic.com/apis/arthur.ai/arthur

# Arthur Scope

Arthur Scope is the monitoring and observability API for production machine learning. An agent can register models, publish metric versions and upload their values, query those metrics over time, and manage the alerts and alert rules that fire when a model drifts or degrades. Workspaces, connectors, and datasets organise the models and the data feeding them, so monitoring stays grouped by team and project.

## For AI agents

Monitor production machine learning models with Arthur Scope: metrics, alerts, alert rules, workspaces, and connectors. Secured with OAuth2.

## Scope

Does not handle model training or feature stores - use it to monitor and alert on production machine learning models only.

## Capabilities

- Register and manage machine learning models
- Publish metric versions and upload their values
- Query model metrics over time
- Create and manage model alerts and alert rules
- Organise models with workspaces and connectors

## Use cases

### Track model performance

Keep an eye on how a deployed model behaves over time. The agent publishes a metric version, uploads its values, and reads the metrics back to spot regressions.

Example prompt: Call POST `/api/v1/models/{model_id}/metrics/versions`, then GET it back to review the values

### Alert on model drift

Catch a degrading model early. The agent defines alert rules on a model and reviews the alerts that fire so a drifting model gets flagged for a human.

Example prompt: Call GET `/api/v1/models/{model_id}/alert_rules`, then GET `/api/v1/models/{model_id}/alerts`

### Organise monitored models

Keep monitoring tidy across teams. The agent groups models into workspaces and wires connectors to the datasets feeding them.

Example prompt: Call GET `/api/v1/workspaces/{workspace_id}` to review a workspace and its models

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| GET | `/api/v1/models/{model_id}/metrics/versions` | Get metric versions for a model |
| POST | `/api/v1/models/{model_id}/metrics/versions` | Create a metric version |
| GET | `/api/v1/models/{model_id}/alerts` | Get model alerts |
| POST | `/api/v1/models/{model_id}/alerts` | Create model alerts |
| GET | `/api/v1/models/{model_id}/alert_rules` | Get model alert rules |
| GET | `/api/v1/connectors/{connector_id}` | Get a connector |
| GET | `/api/v1/workspaces/{workspace_id}` | Get a workspace |
| DELETE | `/api/v1/workspaces/{workspace_id}` | Delete a workspace |

## Key resources

- **Models** — Machine learning models registered for monitoring
- **Metrics** — Metric versions and their uploaded values per model
- **Alerts** — Model alerts and the alert rules that trigger them
- **Workspaces** — Groupings that organise models by team or project
- **Connectors** — Links to the datasets feeding a monitored model

## 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:** 58 / 100
- **Maturity:** Foundational
- **Dimensions:**
  - Foundational Compliance: 65 / 100
  - Developer Experience & Jentic Compatibility: 62 / 100
  - AI-Readiness & Agent Experience: 53 / 100
  - Agent Usability: 43 / 100
  - Security: 90 / 100
  - AI Discoverability: 68 / 100
- **View full report:** https://jentic.com/apis/arthur.ai/arthur/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 Arthur Scope by hand means completing its OAuth2 authorization-code flow, pointing at the platform.arthur.ai host, and shaping model and metric requests yourself. Through Jentic you install once, import Arthur Scope from the Jentic API Directory, store the OAuth2 credentials once, and your agent calls it.
- **Permission scoping:** Arthur Scope puts the model and workspace ids in the URL path (`/api/v1/workspaces/{workspace_id}`), so a rule can pin your agent to one workspace or model and nothing else. You choose which operations it may call, so destructive ones like deleting a workspace are not included unless you add them.
- **Credential handling:** Your Arthur OAuth2 credentials are stored once, encrypted, by your own Jentic One instance and injected at execution time. They never enter the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'get alerts for a model' or 'upload model metrics', and Jentic returns the matching Arthur Scope operation with its input schema so the agent calls the right endpoint without browsing the reference.

## Related APIs

- **New Relic** — Alternative observability API
- **Sentry** — Complementary error monitoring API
- **Comet** — Complementary ML experiment tracking API

## FAQ

### What authentication does Arthur Scope use?

Arthur Scope uses OAuth2 with the authorization-code flow. Through Jentic, those credentials are stored encrypted in your Jentic One instance and injected at execution time, so the raw secrets never enter the agent context.

### What can an agent do with Arthur Scope?

It can register models, publish metric versions and upload their values, query metrics over time, and manage model alerts and alert rules. Workspaces and connectors organise the models and their datasets. Responses are structured so an agent can act on them directly.

### Can I set up model drift alerts through Arthur Scope?

Yes. The API lets an agent create alert rules on a model and read back the alerts that fire, so a drifting or degrading model can be flagged automatically.

### What are the rate limits for Arthur Scope?

Rate limits are not specified in the OpenAPI spec. Check the Arthur documentation for current limits. Through Jentic, retries are handled in the execution layer.

### How many endpoints does Arthur Scope have?

Arthur Scope exposes 202 endpoints covering models, metrics, alerts, workspaces, and connectors.

### Can I limit what my agent is allowed to do with Arthur Scope?

Yes. Jentic One is self-hosted by you, so your own rules decide which Arthur operations your agent may call. Because Arthur puts the model and workspace ids in the URL path, such as `/api/v1/workspaces/{workspace_id}`, a rule can pin the agent to a single workspace and nothing else. You also choose which operations it may call, so destructive ones like DELETE `/api/v1/workspaces/{workspace_id}` to delete a workspace are excluded unless you add them.