Features
EverythingYouNeed,
NothingYouDon't
AI Agents
AI That Actually Understands Your Data
Every AI agent sees your schemas, table relationships, and upstream context automatically. Bring your own model. No black boxes.
Describe your goal: “load my Stripe data and show me revenue by region.” Concierge connects your source, builds the pipeline, runs it, and delivers a dashboard. One conversation, end to end.
Ask ad-hoc questions in plain English and get instant answers. Analyst writes SQL behind the scenes, runs it across all your connected sources, and returns results with inline visualizations like a senior analyst on demand.
SELECT region, AVG(revenue) FROM gold.sales_metrics WHERE quarter = 'Q1 2026' GROUP BY region
Tell it what you need in plain English. AI breaks your request into pipeline steps. Each with validated, production-ready SQL. Review, approve, and deploy in minutes.
Connect any source through conversation. It handles OAuth, discovers your buckets and files, infers schemas, and validates the connection with no forms or config files needed.
Point it at any table and get a full dashboard; widget layout, chart types, filters, and drill-downs chosen automatically. Share with your team in one click.
Every SQL editor has a conversational AI copilot. Ask it to write, refine, or explain your transforms with full context of your schema and upstream tables.
Currently available with Claude, GPT, and Gemini; more models coming soon. You provide your API keys. Your prompts and data stay in your environment.
Set per-workspace custom instructions that shape how every AI assistant behaves; tone, naming conventions, domain context, and guardrails.
Every AI sees your full schema, upstream SQL, and table relationships automatically. No manual context pasting.
AI analyzes your tables and auto-generates validation rules like null checks, range bounds, format patterns, and uniqueness constraints.
Semantic Layer & BI
One Source of Truth for Every Metric
Define metrics once, use them everywhere; dashboards, AI queries, exports. No more conflicting definitions across tools.
Define revenue, churn, ARR, and any custom metric once. Tag them as certified so everyone queries the same number.
Organize dimensions into hierarchies like region → country → city. Drill-down and roll-up just work.
Map joins between tables in your semantic layer. AI uses these relationships to write correct multi-table queries automatically.
Describe the dashboard you need. AI analyzes your tables and generates widgets with the right chart types, filters, and metrics.
Scheduled LLM-narrated insights delivered to email, Slack, or webhooks. Your team gets actionable summaries, not raw data.
Share interactive dashboards with your team. Role-based access controls keep the right people on the right data.
Data Quality
Trust Your Data Before Anyone Sees It
Automatic quality scoring, AI-generated validation rules, and real-time alerts so bad data never reaches your dashboards.
Every table is scored on completeness, accuracy, freshness, validity, and consistency. Quality gates pause pipelines when scores drop below threshold; bad data never reaches downstream.
AI analyzes your data and generates validation rules automatically; null checks, range bounds, format patterns, and custom SQL rules. Mix LLM-generated and hand-crafted.
Get notified when quality drops. Route alerts to Slack, email, or webhooks. Profile any table with one click; distributions, outliers, null rates, and cardinality.
Data Architecture
Your Data, Organized Automatically
Data flows through clean layers such as raw, cleaned, and business-ready. Each step is schema-enforced, auditable, and stored in open formats you own.
Ingest: Raw Data, Untouched
Every record from every source lands here exactly as-is. Full history, zero transforms. Stored with ACID transactions so nothing gets lost.
Clean: Validated & Transformed
AI-generated SQL cleans, deduplicates, and joins your raw tables. Every transform is validated and approved by you before it runs.
Model: Business-Ready Metrics
Aggregated KPIs, dimension tables, and the semantic layer your dashboards query directly. Ready for reporting out of the box.
Open Storage: No Lock-In
Per-workspace data catalogs, time-travel queries, and schema evolution. Your data lives in open formats, portable wherever you go.
Pipeline Canvas
See Your Entire Data Flow
The visual pipeline canvas shows every node in your data pipeline as an interactive graph. Add nodes, preview data at each layer, edit SQL, and connect sources to destinations.
Add source, transform, and destination nodes to your pipeline canvas with a click.
Preview data at any step in your pipeline before deploying to production.
Edit transform SQL directly on the canvas with AI copilot assistance and schema context.
See upstream and downstream dependencies for every node in your pipeline.
Compute & Scheduling
Smart Compute, Automatic Scheduling
Small query? Runs instantly. Big query? Scales up automatically. Pipelines run on schedule in your cloud with no infrastructure to manage.
In-process analytics engine. Runs SQL directly on Parquet and Iceberg files with zero infrastructure. Perfect for datasets under 100GB.
Serverless data warehouse for larger datasets. OptimaFlo routes queries to your warehouse when the data outgrows DuckDB.
Distributed compute for massive-scale workloads. Cluster-level processing for datasets that need it.
Every pipeline runs on Apache Airflow in your BYOC environment. Scheduling, retries, and monitoring built in.
Cron-based scheduling with configurable intervals. Set it once, and your pipelines run on autopilot.
Failed tasks retry automatically. Run backfills across historical date ranges with sequential triggering.
Connectors
Connect, Transform, Export
One-click OAuth or service account authentication. Ingest from any source, export to any destination.
Google's serverless data warehouse
Object storage for files and data lakes
Any REST API with JSON or CSV responses
AWS object storage
AWS data warehouse
Popular relational database
Write processed results to your warehouse for BI tools
Export as Parquet, CSV, or JSON to any GCS bucket
Pipeline completion and schema change notifications via Slack, email, or custom webhooks
Platform & Enterprise
Your Infrastructure, Your Rules
Deploy in your own cloud, own every byte, and prove it to your security team.
Everything runs in your own GCP project. Your data never leaves your environment. OptimaFlo orchestrates, you own the compute.
RBAC, workspace-level permissions, and encryption at rest. Built in from day one. Not sold as an upgrade.
Every pipeline run is tracked with status, timing, and error reporting. Iceberg provides full snapshot history and schema versioning for compliance.
Ready to stop managing infrastructure?
Go from raw data to business dashboards in one conversation.
Now in early beta. Plans from $2,500/mo. Deployed in your cloud. Your data never leaves.
AI-native data platform. From raw data to business dashboards powered by Apache open standards, visual pipeline building, and AI agents that handle the heavy lifting.
© 2026 OptimaFlo. All rights reserved.