--- slug: dashboard/database title: Database in the dashboard kind: reference surface: dashboard summary: Reference for the Database area of the Aetherfy dashboard — the collection inventory, the vector explorer (browse, search, edit a payload, delete or export selected points), the schema and payload index list, and placement — with every label it shows and the afy CLI and API equivalents. sources: - dashboard/src/app/(authenticated)/dashboard/database/page.tsx - dashboard/src/app/(authenticated)/dashboard/database/collections-list.tsx - dashboard/src/app/(authenticated)/dashboard/database/scope.tsx - dashboard/src/app/(authenticated)/dashboard/database/[name]/record.tsx - dashboard/src/app/(authenticated)/dashboard/database/[name]/page.tsx - dashboard/src/app/(authenticated)/dashboard/database/[name]/schema/page.tsx - dashboard/src/app/(authenticated)/dashboard/database/[name]/placement/page.tsx - dashboard/src/components/CollectionWizard.tsx - dashboard/src/components/database/SearchPanel.tsx - dashboard/src/components/database/PointsBrowser.tsx - dashboard/src/components/database/PointEditModal.tsx - dashboard/src/components/database/DeleteConfirmModal.tsx - dashboard/src/components/database/VectorExpandModal.tsx - dashboard/src/components/database/RegionSelector.tsx - dashboard/src/components/database/PayloadIndexList.tsx - dashboard/src/components/SchemaManagement.tsx - dashboard/src/components/SchemaEditor.tsx - dashboard/src/components/SchemaAnalyzer.tsx - dashboard/src/components/regions/CollectionRegionsSettings.tsx - dashboard/src/components/regions/CollectionWorkspaceSettings.tsx - dashboard/src/components/regions/RegionChangeInProgress.tsx - dashboard/src/components/collections/ObservedUsage.tsx - dashboard/src/components/health/DegradedBadge.tsx - dashboard/src/lib/database/payload-indexes.js - dashboard/src/lib/database/page-utils.js - dashboard/src/lib/collections/collection-status-colors.ts - dashboard/src/lib/collections/collectionHealth.js - dashboard/src/lib/health.ts - dashboard/src/lib/proxy/payload-dispatch.js --- # Database in the Aetherfy dashboard ## What the Aetherfy dashboard's Database area is The Database area of the Aetherfy dashboard, at [app.aetherfy.com/dashboard/database](https://app.aetherfy.com/dashboard/database), lists every vector collection on your account and opens one collection at a time. | Address | What it is | |---|---| | `/dashboard/database` | The inventory: each collection's name, dimensions, metric, regions and status | | `/dashboard/database/{name}` | One collection's record. Its root is the **explorer**; the other tabs are **schema** and **placement** | The record's header carries the collection's status, its dimensionality, a reload control and a region selector. On a collection that spans several regions, the selector picks the region the explorer reads from and writes to; it does not move the collection. Placement is changed on the placement tab. ## Aetherfy collection status and health labels | Label | Meaning | |---|---| | `active` | Serving reads and writes | | `creating` | Being provisioned | | `unknown` | No status was recorded | | `DEGRADED` | Healthy enough to use but not where it should be yet; the tooltip says why (below) | | `Replicating to new regions — converging in the background` | A region change is adding or removing regions | | `Moving regions — converging in the background` | The collection is being moved to another workspace | | `Region change failed and rolled back — action needed` | A region change failed and was undone; look at the placement tab | | `Provisioning stalled — collection never finished creating` | The collection has been `creating` for more than 30 minutes | | `NO SUCH COLLECTION` | The list was read and has no collection by this name — it may have been deleted | The dashboard prints a status in capitals. When the collection list cannot be read, the inventory and the record say so instead of claiming you have no collections or that this one does not exist. ## Creating an Aetherfy collection in the dashboard **NEW_COLLECTION** on the inventory opens a wizard: a name (3 to 100 letters, digits, underscores and hyphens) and an optional description, then a template or your own dimensions and distance metric (Cosine, Dot Product or Euclidean), then the regions. On a single-region plan the region is your account's home region; on a multi-region plan you pick from your regions. | Label | Meaning | |---|---| | `OpenAI Embeddings` | 1536 dimensions, cosine | | `OpenAI Large Embeddings` | 3072 dimensions, cosine | | `Sentence Transformers` | 384 dimensions, cosine | | `Custom Configuration` | Your own dimensions and metric | The same thing from a terminal is `afy collections create --size --distance ` ([collections](/cli/vectors)), and over HTTP `POST /api/v1/collections` ([collection routes](/vectors/api)). A collection's name, dimensions and metric are fixed once it exists. ## The Aetherfy vector explorer The explorer is the root of a collection's record. It browses the collection 50 points per page, or runs a query. | Label | Meaning | |---|---| | `QUERY VECTOR` | A vector to search with. Empty means browse | | `FILTER (JSON, optional)` | A Qdrant filter object, applied to a browse or a search | | `SCORE THRESHOLD (optional)` | Drop search results scoring below it. Needs a query vector | | `RUN QUERY` | Run the search, or the filtered browse | | `ALL POINTS` | Browsing the whole collection, a page at a time (`PREV` / `NEXT`) | | `FILTERED POINTS` | Browsing the points a filter matches | | `SEARCH RESULTS` | The nearest points to the query vector, with a `SCORE` column. Not paged | | `NO QUERY RUN` | Nothing has been read yet | | `No results match this query.` | The query matched nothing | | `No vectors in collection. Add the first one to start querying.` | The collection is empty | Each row shows the point's id, a vector preview and a payload preview; the preview leaves out the payload fields Aetherfy reserves, and the point's own view shows the whole payload. | Action | What it does | CLI | REST | |---|---|---|---| | Browse | Pages through the points, without vectors | `afy points list` | `POST /api/v1/collections/{name}/points/scroll` | | Search | The 50 nearest points to a vector | `afy points search` | `POST /api/v1/collections/{name}/points/query` | | View a point | The full vector and payload | `afy points get` | `POST /api/v1/collections/{name}/points/retrieve` | | **NEW_ENTRY** | Inserts one point: id, vector, payload. An existing id is overwritten | — | `PUT /api/v1/collections/{name}/points` | | Edit payload | Replaces one point's whole payload. The vector cannot be edited here | — | `PUT /api/v1/collections/{name}/points/payload` | | Delete a point, or **DELETE SELECTED** | Deletes the points you ticked, on this page, after you type the collection's name | `afy points delete ...` | `POST /api/v1/collections/{name}/points/delete` | | **EXPORT SELECTED** | Downloads the ticked points as a `.jsonl` file, one `{id, vector, payload}` object per line | `afy points get --json` | `POST /api/v1/collections/{name}/points/retrieve` | The commands are documented at [afy vector commands](/cli/vectors) and the routes at [the vector REST API](/vectors/api). Selection covers the current page only and is cleared when the page changes. ## The Aetherfy collection schema tab The `SCHEMA` tab of an Aetherfy collection shows two things: the collection's payload indexes, and its payload schema. | Label | Meaning | |---|---| | `PAYLOAD INDEXES` | The indexed payload fields and their types, as built in the region named beside it | | `No payload indexes: filtering on a field scans every point.` | The collection has no payload index | | `The indexes cannot be read for this collection here: the region that answered did not report them.` | The region did not report them; this says nothing about whether there are any | | `NO SCHEMA DEFINED` | The collection has no payload schema | | `STRICT` | Enforcement: a write that breaks the schema is refused | | `WARN` | Enforcement: a write is checked but accepted | | `OFF` | Enforcement: no validation | The payload index list is read-only. The schema can be created, edited and deleted here, and **ANALYZE DATA** samples the collection's points and suggests a schema. These are the [schema routes](/vectors/api) of the vector API. ## The Aetherfy collection placement tab The `PLACEMENT` tab of an Aetherfy collection changes where its data lives. | Label | Meaning | |---|---| | `PLACEMENT REGIONS` | The regions the collection is replicated to, within its workspace's regions (or your account's, without a workspace). **APPLY** starts a region change | | `REGION CHANGE IN PROGRESS` | A region change is already running, with the region set it is moving to | | `WORKSPACE` | Moves the collection to another workspace, or to none. **MOVE COLLECTION** starts the move. Not offered for the memory SDK's conversation collection, which cannot be moved | | `USED BY` | The agents observed reading or writing the collection, with the deployment, when, and `READ`, `WRITE` or both. Observed, not declared | | `No agent has touched this collection yet.` | No agent was observed using it | Regions and workspace are saved separately: one change at a time. Both are `PATCH /api/v1/collections/{collection_id}` on the control plane — see [moving a collection between workspaces](/agents/api-workspaces) and [Regions and replication](/platform/regions). A region change copies or removes data in the background. **APPLY** first shows what the change will do — the data it copies and deletes, which of your plan's region changes per 30 days it would be, and how much copying is left — and changes nothing until you confirm. A change your plan's storage or region-change limits refuse says why there, and cannot be confirmed. A move changes only which workspace reaches the collection: its data and regions stay where they are, so a destination that does not allow every one of its regions refuses the move, and its aliases are deleted. ## What the Aetherfy dashboard does not do for collections | Not in the dashboard | Where instead | |---|---| | Create or delete a payload index | In code — `create_field_index` (Python) or `createFieldIndex` (JavaScript) — or `afy index create --type ` ([payload indexes](/cli/vectors)) | | Delete points by filter, or more than one page of points at once | `afy points delete --filter …`, which counts what it would delete first | | Delete a collection | `afy collections delete `, or `DELETE /api/v1/collections/{name}`. A collection an agent is using is refused with `COLLECTION_IN_USE` — the `USED BY` list on the placement tab shows which | | Rename a collection, or change its dimensions or metric | Not possible in Aetherfy: create a new collection and copy the points, or give it a second name with a [collection alias](/vectors/api) | | Edit a point's vector | Insert the same id again with **NEW_ENTRY**, or upsert from code | | Change a collection's description | `PATCH /api/v1/collections/{name}` with `description` |