Install Jentic One Beta
Jentic One is a self-hosted execution layer for AI agents. It lets your agent call the Dagster External Assets API, or any other public or private API you need. You set the rules, the agent never sees your credentials, and every call is logged.
Two steps, two machines. Install the instance in a safe environment, then register your agent from wherever it runs.
Step 1: Jentic One Host machine
# On the machine that will host your Jentic One instance:
curl -fsSL "https://jentic.com/install.sh?src=apis&api=%2Fapis%2Fdagster.io%2Fdagster" | shStep 2: Agent machine
# On the machine where your agent runs (keep this separate from the instance):
curl -fsSL "https://jentic.com/install.sh?src=apis&api=%2Fapis%2Fdagster.io%2Fdagster" | sh
jentic register # connects your agent to your Jentic One instanceJentic One is in public beta. The setup above keeps your agent separate from the instance, which is what you want before using real credentials: an agent running as the same OS user as Jentic One can read its stored keys directly. Just evaluating? A single local install is fine to start. See the secure deployment guide for the tiers.
What an agent can do with Dagster External Assets API.
Report an asset materialization when an external job writes data
Report the result of an asset check against a materialized asset
Report an asset observation to record metadata without a materialization
Attach a partition key to a reported asset event
Record a data version and custom metadata on an asset event for lineage
GET STARTED
Patterns agents use Dagster External Assets API for, with concrete tasks.
★ External Pipeline Materialization Reporting
When a warehouse job or notebook outside Dagster finishes writing a table, an agent calls POST /report_asset_materialization/ with the asset_key and a metadata payload. Dagster Cloud then shows the asset as freshly materialized even though Dagster never ran the job, keeping the lineage graph honest.
After an external job writes the orders table, report a materialization for its asset_key with row-count metadata
Data-Quality Check Reporting
A data-quality test that runs in an external system posts its pass or fail verdict with POST /report_asset_check/. This surfaces the check status next to the asset in Dagster Cloud so freshness and quality signals stay in one place rather than in a separate dashboard.
When the nightly null-rate test finishes, report its pass or fail result against the asset it validated
Metadata Observation Without Materialization
To record that an asset was inspected without rewriting it, an agent sends POST /report_asset_observation/ with updated metadata such as size or last-modified time. This keeps observability metrics current for slowly changing sources that are not recomputed on every run.
Report an observation carrying the current file size and last-modified metadata for a source asset
Partition-Aware External Reporting
For partitioned assets, each report includes a partition key so Dagster tracks which slice was updated. An agent processing a daily partition reports the materialization for that specific partition, keeping the partition status map accurate for backfills and freshness checks.
Report a materialization for the 2026-09-14 daily partition of the events asset
3 endpoints — the dagster external assets api reports events about assets that live outside dagster into dagster cloud, so an orchestration graph stays aware of work done elsewhere.
METHOD
PATH
DESCRIPTION
/report_asset_materialization/
Report a materialization event for an external asset
/report_asset_check/
Report the result of an asset check
/report_asset_observation/
Report an observation with metadata for an asset
/report_asset_materialization/
Report a materialization event for an external asset
/report_asset_check/
Report the result of an asset check
/report_asset_observation/
Report an observation with metadata for an asset
What agents get from Jentic-routed access to this vendor.
Setup
Calling the Dagster External Assets API by hand means building the `Dagster-Cloud-Api-Token` header, pointing at your organization's Dagster Cloud deployment, and shaping each report body yourself. Through Jentic you install once, import the API from the API Directory, store the token once, and your agent reports events.
Permission scoping
This API reports events through three operations and takes its target asset in the request body, so scope it by operation: allow only the reports the agent should send, for example materializations and observations, and leave check reporting out unless you add it.
Credential isolation
Your Dagster Cloud agent token 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.
Intent-based discovery
Agents search Jentic by intent such as 'report an asset materialization' or 'record an asset check result', and Jentic returns the matching Dagster operation with its input schema so the agent reports against the right endpoint.
Alternatives and complements available in the Jentic catalogue.
Specific to using Dagster External Assets API through Jentic.
What does the Dagster External Assets API do?
It reports events about assets that live outside Dagster into Dagster Cloud. You can report a materialization when an external job writes data, an asset check result from a data-quality test, or an observation that records metadata without a materialization, each with an optional data version and partition.
What authentication does the Dagster External Assets API use?
Requests carry an agent token in the `Dagster-Cloud-Api-Token` header, scoped to your organization and deployment. Through Jentic the token is held by your own instance and supplied at execution time, so the agent never reads the raw value.
Can I report a materialization for an external job?
Yes. POST /report_asset_materialization/ records that an asset was written outside Dagster, including its `asset_key`, a `data_version`, a `partition`, and any `metadata` you attach. Dagster Cloud then treats the asset as freshly materialized.
How do I record a data-quality result?
Send the verdict with POST /report_asset_check/ so the check status appears next to the asset in Dagster Cloud. Use POST /report_asset_observation/ instead when you only want to update metadata without marking a materialization.
What are the rate limits for the Dagster External Assets API?
The OpenAPI specification does not declare rate limits. Report events as work completes rather than in tight loops, and back off if a request returns an error response; confirm current limits in the Dagster Cloud documentation.
Do I need an MCP server to use the Dagster External Assets API with an agent?
You do not. Jentic connects the API to your agent directly from the API Directory: import it, store your token once, and the agent calls the reporting operations without a separate server to run.
Can I limit what my agent is allowed to do with the Dagster External Assets API?
Yes. Because you run Jentic One yourself, you pick which of the three reporting operations the agent may call. A reporting agent can be given materialization and observation reporting while asset-check reporting stays out of its allowed set, and the token stays with your instance so the agent acts only within that choice.
For Agents
Report asset materializations, check results, and observations for external assets into Dagster Cloud, each with optional metadata, data version, and partition.
Use for: I need to report an asset materialization to Dagster Cloud, Record an asset check result after a data-quality test runs, Send an asset observation with the latest row count, Tag a reported materialization with its partition key
Not supported: Does not run pipelines, manage deployments, or query historical asset events. Use it to report materializations, asset-check results, and observations for external assets only.
The Dagster External Assets API reports events about assets that live outside Dagster into Dagster Cloud, so an orchestration graph stays aware of work done elsewhere. It records materializations when an external job writes a table or file, check results when a data-quality test runs, and observations that capture metadata without a materialization. Each report can carry a data version, a partition key, and arbitrary metadata, which keeps lineage and freshness accurate for assets Dagster does not compute itself. Data and platform teams use it to fold external pipelines, notebooks, and warehouse jobs into a single Dagster asset view.