Give your data a goal.It runs the whole analysis.
The agentic BI platform for teams. Your whole company gets an AI analyst that runs the investigation on your own databases — queries, dashboards, monitoring. Every number cited, never guessed.
No credit card · 14-day team trial · Read-only access to your data
Margin analysis
sales_data.csv · example data
Question
Why is our profit margin going down?
Gross margin by product
%
Root cause
Sneakers are returned 3× more than any other product — and every return wipes out the sale.
One analyst for every BI job
Deep analysis, dashboards, and quick answers — one place, one source of truth.
Dashboards that build themselves
- Ask “build me a dashboard” — it picks the KPIs and charts.
- Live, not snapshots — every widget re-queries on each view.
- Click any bar to cross-filter the whole board.
Revenue
$6.3M
+8.2%
Gross margin
26.4%
−1.2pt
Return rate
6.7%
+0.4pt
Revenue by region
Revenue trend
A quick answer, instantly
- Ask in plain English — no SQL.
- Get the number, the change, and the why in one reply.
- See the exact SQL behind every reply.
Total sales last month: $284,500
That's +12.4% compared to the previous month. Your best performing region was the West Coast.
An analyst that does the work — not just a chatbox
Agentic analysis
Give it a goal, not just a question. It plans a multi-step investigation, runs it, and drills into whatever it finds.
Live dashboards, built for you
Ask for a dashboard — it picks the KPIs and charts, re-querying live on every view. Click any bar to cross-filter.
Proactive monitoring
Put any finding on watch. Get alerted when a metric crosses a threshold, plus a digest of what changed.
Adaptive thinking
Watch it reason step by step — it chases outliers, then stops when the question is answered.
Governed & trustworthy
Define a metric once; every analysis uses the same formula. If the data isn't there, it says so.
Connect anything
PostgreSQL (including Supabase, Neon and Amazon RDS), MySQL, BigQuery, Snowflake, Google Sheets or CSV. Read-only, encrypted, connected in 60 seconds.
100%
of numbers cited to a query
0
write access — SELECT-only
6
data sources supported
60s
to connect and start asking
Works with your data sources
Read-only, credentials encrypted. See all integrations →
Put your data to work, programmatically
Ask questions, run read-only queries, and pull charts and dashboards from your own backends, notebooks, Slack bots, and automations — with a simple Bearer-token REST API.
- Bearer-token auth — create and revoke keys in Settings
- Read-only and safety-checked, with per-plan rate limits
- Same answers, SQL, rows, and chart config as the app
curl https://www.schemachat.co/api/v1/chat \
-H "Authorization: Bearer $SCHEMACHAT_API_KEY" \
-H "Content-Type: application/json" \
-d '{ "question": "Top customers by revenue last month?" }'
# → { "answer": "...", "sqlQuery": "...",
# "queryResult": [...], "chartConfig": {...} }Your AI and your dashboards can't disagree
Both read from metrics you define once. Here's how the numbers stay honest and your data stays safe.
Every number is cited — or it says no
Every figure comes from a query that actually ran, with the step behind it cited. If the data isn't there, it says so — it never fabricates.
Governed metrics, computed one way
Define Average Order Value once; every analysis and dashboard uses that exact formula — so two reports never contradict each other.
Read-only by design
All AI queries are SELECT-only. We physically cannot modify or delete your data — enforced at the SQL parser layer.
Credentials encrypted at rest
Database credentials are encrypted with AES-256-GCM before being stored. Decryption only happens server-side during a live query.
Built for teams
Roll it out to the whole company, with the controls IT expects
Roles that match your org
Admins manage people, billing and data sources; members build and ask; viewers see and ask but can never change anything — enforced on the server.
Your company, not a stranger's
Colleagues with a verified company email can join your workspace automatically, within the seats you have. Competitors can't claim your domain.
Accountability built in
An activity log records who invited whom, who connected or deleted a data source, and which API keys were created — downloadable as CSV.
Security you can enforce
Require two-factor authentication for every member, export or delete your workspace's data yourself, and see exactly what reaches the AI.
One set of numbers
Shared metrics and business terms mean finance, sales and operations all get the same answer to the same question.
Usage you can justify
See who is asking what and how often, per person, so you can show the value before you renew.
Honest about what we don't have yet: no SOC 2 or single sign-on today. Read the security page or talk to us.
Put an AI analyst to work for your whole team
Start a 14-day team trial: connect your own database read-only, run deep analyses and build dashboards. Roll out to the company when you are ready.
Start free team trial Talk to salesFrequently asked questions
- What databases does SchemaChat work with?
- PostgreSQL (including Supabase, Neon, Amazon RDS and Aurora), MySQL, BigQuery, Snowflake, Google Sheets shared by link, and CSV/Excel uploads. Redshift, CockroachDB and TiDB are in beta.
- Can it change or delete my data?
- No. Access is read-only, and queries that write or alter data are refused.
- How do I know an answer is right?
- Every answer that runs a query shows its SQL and a “How this was computed” receipt. We also test answers on every release against 115 questions whose correct answers are computed independently — and we still recommend checking numbers that matter.
- What happens when my data can't answer the question?
- It says so instead of guessing, and a multi-step analysis lists anything it could not cover.
- What is the difference between Ask and Analyze?
- Ask answers one question. Analyze takes a goal, proposes a plan you can edit, runs several queries, checks each one and writes a summary that cites its steps.
- Do I need to know SQL?
- No. The SQL is shown so someone who knows it can check it, but you don't have to read it.
- Do I need a data model or semantic layer first?
- No. SchemaChat reads your schema when you connect. You can add governed metrics later if you want fixed definitions.
- Who can see what?
- Workspaces have admin, member and viewer roles, an audit log, and optional required two-factor sign-in for everyone.
- Is there a free trial?
- Yes — 14 days for a team, with no credit card.
- Does it have SOC 2 or single sign-on?
- Not yet. The security page lists what we do have, and you can request our security questionnaire from sales.