Back to the platform overview
Roadmap
What lavalake does today and what comes next
Version 0.201.0, checked on October 17, 2026. Available means shipped. PostgreSQL, MySQL, SQL Server, Snowflake, JDBC and Python have been tested against real systems; SAP, CRM, DATEV, Tableau, Power BI and ODBC/Excel still await acceptance. Features requiring Compose or Kubernetes state their deployment scope.
Status
Click to filter
AI and assistant
The layer this is about: questions to your own data, under the rights of whoever asks.
- AvailableAssistant with curated tools, confirmation before every change
- AvailableMCP server that passes through the user's own permissions
- AvailableRegistry for third-party MCP servers in the console
- AvailableChoice of language model per installation, self-hosted included
- AvailableSigning in with your own Claude subscription as agent mode
- AvailableDashboards fully editable through the assistant
- AvailablePyro displays charts in chat and adds them to dashboards
- AvailableConversations persist, with a list and a self-written title
- AvailableAssistant queries curated metrics through stored definitions
- AvailableAssistant creates files and exports query results as CSV
- AvailableAssistant answers questions about platform load
- AvailableQuestions in your own words against curated metrics with Pyro
- AvailableAI in SQL: summarize, classify, translate and extract
- AvailableAzure OpenAI, AWS Bedrock and Google Vertex AI as providers
Lakehouse and queries
Open formats on your own storage, so the data stays readable without us.
- AvailableApache Iceberg on S3-compatible object storage
- AvailableVersioned catalog with time travel and branching
- AvailableTrino as the query engine, SQL worksheet in the console
- AvailableTable streams for net changes and Snowflake SQL translation in the editor
- AvailableData browser across catalog, schema and table
- AvailableWorking copies of the data, reversibly resettable to an earlier state
- AvailableS, M and L workload classes per role in the Compose stack
- AvailableScheduled Iceberg maintenance: compact files, expire snapshots and remove orphan files
- AvailableQuery tuning: detect missing physical layout and apply it in one click
- AvailableTable statistics as the fourth step of Iceberg maintenance
- AvailableScaling out with additional worker nodes
- AvailableWarehouse quotas; dedicated Kubernetes clusters, real-cluster KEDA acceptance pending
- AvailableMaterialized metrics with a watermark instead of full recomputation
- AvailableSaved views per data pool, executed under the caller's own identity
- AvailableSQL editor with syntax highlighting and suggestions from your own catalog
- AvailableWorkload separation by purpose, so a dashboard cannot starve an import
- AvailableFilesystem cache for hot Parquet blocks on the compute nodes
Data ingestion
Four routes into the lakehouse, each with an assistant instead of a config file.
- AvailableImport assistant with modelling and cleanup before loading
- AvailableConnectors for PostgreSQL, MySQL and Microsoft SQL Server
- AvailableFiles, documents and a file crawler as sources
- AvailableUnified file upload for Excel (XLSX), CSV, PDF, DOCX, TXT and Markdown
- AvailableThird-party MCP servers as a data source
- AvailableSAP Business One as a source through the Service Layer
- AvailableSAP S/4HANA through the gateway's OData service
- AvailableDATEV files recognised as DATEV
- AvailableImports keep running on the server, even after you sign out
- AvailableContinuous ingestion instead of full load runs
- AvailableInbox: new files in object storage are imported automatically
- AvailableDelta load on a schedule: only new rows, using a watermark column
- AvailableLoad modes new, append and replace — through the import API as well
- AvailableSource import without a row cap, with per-table progress
- AvailableGuided “add data” flow from the source through to the schedule
- AvailablePostgreSQL CDC through logical replication: inserts, updates and deletes; Compose
- AvailableRecurring schedules: analyses that run by themselves
- Availablelavaflow: build ETL pipelines by clicking instead of typing
- Availablelavaflow: branching, quality checks, HTTP, forecasts and Pyro transformations
- AvailableCRM sources: Microsoft Dynamics 365, HubSpot and Salesforce
- AvailableOAuth connections for external SaaS sources, with no password in the config
- AvailableSchedules can be chained: the next starts after a successful run
- AvailableAssisted data modelling inside the import assistant
Search, metrics and BI
One definition per metric – the same number in chat, dashboard and BI tool.
- AvailableHybrid search combining vector and keyword retrieval
- AvailableEmbeddings from a local model, no external service
- AvailableCurated metrics and semantic views
- AvailableDashboards that compute under each viewer's own rights
- AvailableTableau reads the lakehouse through a read-only feed
- AvailablePower BI through a bundled Power Query connector
- AvailableAlerts against thresholds, reports stored as snapshots
- AvailableDashboards with an edit mode, a filter bar and more than a dozen tile types
- AvailableSchedule reports from dashboards and SQL queries and send them by email
- AvailableJDBC and ODBC access to Trino, for Excel and your own tools
- AvailableData product catalog: publish and find curated datasets internally
- AvailableFederated live queries against external systems, tightly bounded
Permissions and evidence
Nobody gets around the permissions – the agent included.
- AvailableDelegated identity through Keycloak, no service account
- AvailableRole-based row and column policies in Trino (Compose)
- AvailableTamper-evident audit log as a hash chain
- AvailableLineage and access history per table
- AvailableUser and role administration in the console
- AvailablePolicies report whether they actually take effect
- AvailableTable and column classification and tags, suggestions based only on names and types
- AvailableSign-in through Microsoft Entra ID (SSO)
- AvailableMachine access for scripts and pipelines, as its own principal in the audit log
- AvailableSelf-service “My account”: password, sessions, two-factor
- AvailableRole hierarchy and per-object ownership
- AvailablePersonal-data access, deletion and retention workflows
- AvailableSensitive data classification and a record of processing activities
- AvailableDashboard visibility by role
- AvailableLong-lived, revocable access tokens per connected AI system
Operations
Reference deployment on one machine with Compose. Installation through Ansible, Kubernetes as a separate deployment path.
- AvailableCompose stack for a single machine, Helm chart for Kubernetes
- AvailableMonitoring and audit inspection in the console
- AvailableConsumption transparency: who consumes how much, and for what
- AvailableAnsible playbook for installation and updates on Debian/Ubuntu; full run on a fresh VM still to verify
- AvailableDaily database and configuration backup in the Compose stack with visible status
- AvailableSync AI list prices on demand, preserving contract prices
- AvailablePublic access over HTTPS under one shared hostname
- AvailableAnalytics board: queries, data, compute time, AI cost and license status
- AvailableLoad cockpit with traffic lights and a query time series
- AvailableLicense model, verified offline and without technical enforcement
- In progressKubernetes HA profile available; live-cluster and failure acceptance outstanding
- AvailableHealth, alerts and logs in the console; /metrics for your own Prometheus
- AvailableOptional internal TLS layer for documented service connections
- AvailableSecret files for Compose; Vault/External Secrets for Kubernetes
- AvailablePoint-in-time recovery, with backups to a second location
- AvailableOffline bundle and air-gapped Compose deployment with local models
- AvailableDaily quotas per user for compute time and data read
- AvailableAI usage per user and model, costs from a price table and a daily token budget
- AvailableCustomer installation through Ansible, with no manual steps on the machine
- AvailableMigration toolkit for Snowflake: inventory, translate, verify
Deliberately not planned
Large product areas we are not taking on. Anyone who needs them is better served by a cloud warehouse – saying so is more honest than leaving it open.
- Not plannedData marketplace and listings
- Not plannedSecure data sharing between organisations
- Not plannedReader accounts for external parties without their own installation
- Not plannedCross-region replication
- Not plannedRunning application code inside the database
- Not plannedHosting third-party applications on the platform
Something missing, or something urgent?
The order comes from customer questions, not from a drawing board. Tell us what you need – that moves the list more than a competitive comparison does.
Arrange a conversation