See everything you run.
Change it with confidence.
Discovery Accelerator is a local-first engineering workbench that maps your databases, warehouses, cloud, and infrastructure — so every engineer knows what exists, how it connects, whether it's healthy, and what's safe to change.
The questions that stall every change.
Engineers lose hours answering the same things before they can safely touch anything. Discovery Accelerator keeps the answers current.
Discover. Diagnose. Document. Accelerate.
Four stages that turn an unfamiliar estate into something you can reason about and safely operate.
Discover
Reuse existing az login, AWS profiles, and gcloud ADC to inventory databases, warehouses, lakehouses, cloud, and infrastructure — read-only, no agents.
Diagnose
Connection diagnostics, environment analysis, health, and drift — pinpoint what's misconfigured before it becomes an incident.
Document
Automatic ERDs, data dictionaries, relationships, and lineage — living documentation that stays current with the estate.
Accelerate
AI-assisted schema understanding and safe SQL generation — metadata-only, always read-only, and always previewed before it executes.
One workbench for the whole estate.
From transactional databases to cloud infrastructure and the code that provisions it.
Built to be trusted in regulated environments.
Security isn't a feature bolted on — it's the default posture. Nothing leaves your machine unless you send it there.
Read the trust postureLocal-first by design
The catalog and every credential stay on your machine. There is no vendor cloud collecting your metadata.
Read-only by default
A tokenizer-based safety guard permits only read statements. Every generated query is previewed before it can run.
Credentials reused, never scraped
It uses the sessions you already trust — az login, AWS profiles, gcloud ADC. Secrets are never logged or serialized to disk.
AI sees metadata only
The assistant receives table and column names, types, and relationship flags — never the contents of a single row. Bring your own model: Azure AI Foundry, Anthropic, OpenAI, or a local one.
Own it. Once. For life.
Every paid edition is a one-time, perpetual license — buy once, use forever, no subscription. Bring your own AI, or top up prepaid Managed AI credits.
The Founding Engineering Council
The first 25 organizations accepted don't get a discount — they get a seat at the table, and a lifetime Professional license to match.
Apply for a founding seatWorth far more than “free.”
- Lifetime Professional license
- Every future Professional upgrade
- Founding customer badge
- Direct influence on the roadmap
- Private Discord / Teams channel
- Quarterly roadmap calls with engineering
For engineers exploring their environment — local, private, unlimited.
- All 10+ database & warehouse engines
- Azure, AWS & GCP discovery
- ERDs, catalog, diagnostics & docs
- Safe SQL & query history
- Bring your own AI & cloud credentials
- Unlimited use · community support
For consultants, architects, and DBAs who live in this daily.
- Everything in Community
- Workspace projects & saved sessions
- Report generation & advanced ERDs
- Recon Center & AI prompt library
- Schema compare & drift detection
- Priority updates · bring your own AI
For small IT teams standardizing on one source of truth.
- Everything in Professional
- Shared projects & catalog
- Team reports & discovery history
- Role permissions & central config
- Export templates
- Email support
For large organizations with governance and deployment needs.
- Everything in Team
- SSO / Entra & audit logging
- API, automation & CLI
- Air-gapped & offline deployment
- Custom connectors
- Priority engineering & onboarding
Bring your own AI, free — or add Managed AI credits $50 / $100 / $250, prepaid, no subscription. Prefer we host it? Discovery Cloud — Starter $995/yr · Business $2,995/yr · Enterprise custom.
What engineering and security ask first.
Does my data ever leave my machine?
No. Discovery Accelerator is local-first — the catalog, credentials, and results stay on the workstation. Only the metadata you explicitly send to your chosen AI model leaves, and it never includes row data.
Can it write to my databases?
Not by default. A tokenizer-based safety guard allows only read statements, enforces read-only sessions at the connection layer, and previews every query before it runs.
Which AI models does it use?
Bring your own — Azure AI Foundry, Anthropic, OpenAI, or a local model such as Ollama. If none is configured it falls back to an offline mock, so the app always works.
How is it licensed?
Community is free forever. Professional, Team, and Enterprise are one-time, perpetual licenses — buy once, own it for life, no subscription. Bring your own AI, or add prepaid Managed AI credits. Enterprise adds SSO, APIs, and air-gapped deployment.
Do you support air-gapped or on-prem?
Yes, on the Enterprise plan — including offline installation, an internal AI endpoint, and no outbound connectivity requirement.
What does it run on?
A native Windows 11 desktop workbench today, with the same engine reachable through a CLI for automation and CI.
Start mapping your estate today.
Install on one machine, connect the sessions you already have, and see your first estate map in minutes — read-only, from the first click.