Cloud changelog

Updates to your hosted little big brain workspace. For database behavior and developer tools, see the Engine changelog.

Entries describe the terms and behavior at release time. For today’s prices, see Usage & billing.

October 2026

8 October: A dashboard on Home

Home now shows the selected graph at a glance. The top panel gives its state and its size: instances, classes, statements and commits. A chart shows the month’s model work per day, by job.

Each feature has a tile with one live figure. Examples are searchable classes, open suggestions and the share of checked answers that were right. The largest classes and the newest background work close the page.

8 October: Upload files from Integrations

Integrations now holds every way data comes in. File upload sits beside the connectors and takes Turtle, N-Triples or NDJSON. It works on every plan, also where connectors are off. The separate Data page is gone, and its old link opens Integrations.

8 October: A guided first run

A new workspace opens on four steps: environment, graph, data and first questions. Each step offers its options, such as a source, a starter or a file. A workspace part way through sees the same steps above its dashboard.

When a feature cannot run, the page says why and what to do next. Reasons include a step still to do, a Cloud feature on Free, or a Free limit. On Free, the sidebar shows your storage against the 30 GiB limit.

8 October: Clearer navigation

The sidebar groups the pages by the work: Graph, Sources, AI, Operate and Developers. Each page names its group above its title. Status shows a dot while the graph updates or needs attention. Knowledge graph shows the count of open suggestions.

The command palette (⌘K) uses the same groups. It also finds a page by what it does. “sparql” finds Query, and “api key” finds Access keys. A Docs link sits at the foot of the sidebar.

4 October: Free includes 30 GiB

Free now includes 30 GiB of data, indexes and retained history, up from 10 GiB. Every Free account gets the new allowance, including accounts that started on Free before this change. The other Free terms stay the same. See Usage & billing.

4 October: A Rerank switch beside the search box

In Knowledge graph, each search box has a Rerank switch. It turns the rerank on or off for the graph. Jev then puts first the hits that answer your words. A line above the hits says their order, and each hit shows its relevance. Models has a new Search rerank page with the month’s ranked hits and each graph’s setting. See Search by meaning.

2 October: Start a graph from a starter

A starter is a ready ontology for a domain: CRM, Documents or Work management. Each one holds classes, relations and the questions it answers. Knowledge graph has a new Starters tab with each starter and its status on the graph. Apply first shows what the starter adds. It never removes or changes what the graph holds. Create graph can start from one or more starters. A synced source starts from its domain’s starter: HubSpot from CRM, Linear from Work management, SharePoint from Documents. See Ontology starters.

2 October: Search by meaning on Free

Free now includes search by meaning. Turn it on for a class in Knowledge graph. The platform embeds the records with its managed model, at no charge. The vectors count toward the 10 GiB. See Search by meaning.

1 October: A free plan

Start on Free with no card. Free includes one environment and 10 GiB of data, indexes and retained history. It has no expiry and no automatic charges. It allows 300 requests per rolling minute, reads and writes together. Choose Start on Free when you set up a workspace. Upgrade to Cloud when you need more; your data and keys stay. Managed models require Cloud. See Usage & billing.

1 October: Choose how a graph starts

Create graph now asks how to start. Import a file, write from code, or design the ontology. Where integrations run, you can also sync a source. LBB then proposes the ontology from the source’s records. The console opens the page for your choice. The other ways stay open later.

September 2026

30 September: See how a query runs

Turn on Profile in Query to see how a query ran. The profile shows the time per stage, the blocks and bytes read, and the join order the planner chose, with its estimate for each pattern. See See how a query ran.

Query planner, under Operate, shows the statistics the planner uses for a graph: the triple count of each predicate and the indexes that help it plan. It can build the value-order index. The profiles you ran in the tab are listed there too.

28 September: Strong reads by default in Query

Query now runs strong reads by default, so a query sees your latest writes. A long queue of unindexed commits makes a query slower, not refused.

28 September: Search by meaning for every account

Search by meaning is now on for every account. Turn it on for a class in Knowledge graph.

28 September: A shorter Billing page

Billing has less text. Storage-only plans show one Storage table: what each environment stores, its search vectors, and the cost per month. The page and Status also load faster.

28 September: Rules you can read

The Validation tab shows each SHACL rule as a short sentence, grouped by class. An example: “Each Service has at least 1 tier”. Each rule shows how many entities break it. Open it to see them by name.

A class card and a property page list the rules that apply to them. A red dot marks a class or property with failing entities. On Status, the current search classes fold into one list with a filter.

28 September: Knowledge graph in the sidebar

Ontology is now Knowledge graph. Its first tab, Browse, is now Ontology. Status, Environments and Graphs have sidebar items. Data left the sidebar. Open it from Home, Status or Graphs.

27 September: See the work on a graph

A dot beside the graph name shows the graph’s state. Select it to open Status. Status shows how far the newest commit has reached: written, indexed, and searchable. It lists running work: index updates, search embeddings, and imports. Recent finished work stays below.

26 September: Browse the graph in Ontology

Explorer is now the Browse tab of Ontology. The class tree sits on the left. A class sits on the right with its links on either side. An instance opens as a record, with its values and relationships. Old Explorer links open the same view.

19 September: Storage-only Cloud pricing

New Cloud subscriptions pay a €19 monthly minimum and €0.45 per GiB-month, including applicable VAT, or $21 and $0.49 per GiB-month plus applicable tax. Storage includes indexes and retained history. Queries, imports, and index builds are included, with no separate usage charge.

The minimum is prepaid and covers storage for the billing period; storage above it is invoiced after the period ends. Existing resource and query-unit subscriptions keep their accepted prices and schedule until an agreed change. See current prices and examples.

10 September: Resource pricing and VAT-inclusive subscriptions

New Cloud subscriptions pay €90 including applicable VAT or $100 plus applicable tax upfront, then monthly on the signup date. Usage draws from that monthly minimum, with account-wide storage tiers and resource work across queries, imports and index builds. Resource plans have no separate query-count charge.

The signup screen shows the accepted prices and first-payment progress. Existing subscriptions retain their current rates and schedule. See Usage & billing for the resource-hour formula, spending controls and payment recovery.

4 September: Clearer navigation and account usage

Graphs and environments are accessible from the switchers at the top of the sidebar. Billing and Settings are in the account menu at the bottom.

Billing shows storage and query-unit usage by environment and for the account, with the rate and subtotal for each. You can review your plan, credits and their expiry dates, charges, and latest invoice in one place. Account owners and admins can review plan terms before continuing to payment setup when it is available for their account.

See Usage & billing for current plans and availability.

3 September: Shared cache is included

Caching is managed automatically and has no separate charge. Account usage includes storage and query units.

2 September: Account-wide billing visibility

The Billing page brings together usage across your environments, account credits, and invoice status. Owners and admins can manage payment details; members can view account usage.

August 2026

31 August: Typed properties in the Ontology workbench

Imported ontology data properties appear as fields with their native value types. This makes text, numbers, booleans, and dates easier to distinguish from relationships between classes.

28 August: Browse the ontology hierarchy

The Ontology screen shows each class with its direct parents and children. You can browse instances through the class hierarchy and inspect how classes relate to one another.

23 August: Explore RDF data through classes and relationships

Explorer connects observed classes through their relationships. The instance inspector shows RDF fields and grouped relationships, with one-hop expansion and traversal between related entities. When data is still being indexed, the console provides clearer retry guidance.

22 August: Faster class browsing and simpler administration

Opening a populated class loads a sample of its instances, keeping Explorer responsive on larger datasets. Graph cards are directly selectable, and owners and admins can manage members and access keys.

July 2026

27 July: Sign-in during large imports

Account sign-in stays responsive while data is being indexed.

24 July: Navigate between classes and instances

Explorer lets you move from ontology classes to their instances and back, with related entities grouped by relationship.

21 July: Choose read consistency in the console

Query and Explorer let you choose read consistency when you need newly written data to be visible. See Core concepts.

20 July: Revoke access keys

Revoking a stack access key takes effect immediately.

16 July: Stack endpoints and guided onboarding

Each stack has its own endpoint, ready to copy into your application. You can create named access keys, and a first-run checklist guides you from stack creation to your first query.

See Endpoints & base URLs and Authentication & keys.

10 July: Import Turtle files

The console accepts RDF Turtle files directly.

5 July: Browse with your account session

Your signed-in console session can read stack data without entering an access key in Explorer.

2 July: Redesigned console

The console adds reorganized navigation, a visual ontology editor, and guided setup for a new stack.

June 2026

19 June: Connect MCP with your account

Connect an MCP client using account sign-in. Agents can use the stack’s query and write tools without handling raw access keys.

See MCP for agents.