Contexts
Contexts represent the devices, users, and other entities that evaluate your feature flags. Every time an SDK evaluates a flag, it reports the evaluation context — who or what requested the flag value, along with any attributes attached to that entity.
The Contexts dashboard gives you visibility into every entity that has interacted with your flags: what they evaluated, when, and which variations they received.
Context kinds
Each context has a kind that describes the type of entity it represents:
| Kind | Description |
|---|---|
| user | An individual user of your application |
| device | A specific device (phone, tablet, browser instance) |
| organization | A company, team, or account that groups multiple users |
| service | A backend service or microservice evaluating flags server-side |
| environment | A deployment environment (production, staging, development) |
The kind determines how the context icon appears in the dashboard (each kind has a distinct color-coded icon) and lets you filter the contexts list.
How contexts are created
Contexts are created automatically when your SDK evaluates flags. When you call OpenFeature.setContext() and flags are evaluated, the SDK reports the evaluation context back to AppDispatch. No manual setup is needed — as soon as devices start evaluating flags, their contexts appear in the dashboard.
For SDK setup details, see Setting evaluation context. For how evaluations are reported, see OpenFeature integration.
You can also create contexts manually from the dashboard.
Contexts list view
The contexts list shows all known evaluation contexts with:
| Column | Description |
|---|---|
| Kind icon | Color-coded icon indicating the context kind |
| Targeting key | The unique identifier for this context within its kind |
| Kind badge | Label showing user, device, organization, service, or environment |
| Name | Optional display name |
| Attribute chips | Up to 3 attribute key-value pairs displayed inline, with an overflow indicator if more exist |
| Evaluation count | Total number of flag evaluations for this context |
| Last seen | When this context last evaluated a flag |
Results are paginated at 50 contexts per page.
Filtering
- Search — Filter by targeting key, name, or attribute value.
- Kind filter — Dropdown to show only contexts of a specific kind.
Context detail view
Click any context to open its detail dialog. The detail view shows:
Summary
- Targeting key, kind, and name
- Total evaluations, first seen, and last seen timestamps
Attributes table
All attributes reported by the SDK for this context, displayed as key-value pairs.
Flag evaluation history
A per-flag breakdown of every flag this context has evaluated:
| Column | Description |
|---|---|
| Flag name | Display name of the flag |
| Key | The flag key |
| Channel | Which channel the evaluation occurred in |
| Variation | The variation value the context received |
| Eval count | Number of times this context evaluated this flag |
| Last evaluated | When the most recent evaluation occurred |
Creating contexts manually
While contexts are typically created automatically via SDK evaluations, you can create them manually from the dashboard. Click Create Context and fill in:
| Field | Description |
|---|---|
| Kind | Select from a visual grid of the 5 context kinds |
| Targeting key | Unique identifier for this context |
| Name | Optional display name |
| Attributes | Dynamic key-value pairs — add as many as needed |
Manually created contexts are useful for testing flag targeting rules before devices start reporting in.
Deleting contexts
To delete a context, click the delete button on the context row or in the detail view. Deleting a context removes all evaluation history for that entity. This action cannot be undone.