THE OPENDASH PRODUCT

One analyst that
shows its work.

Ask your database in plain English. Every answer comes with the SQL and rows behind it. Every correction can make the next answer better.

Book a demo
Pair, the OpenDash mascot: a lavender plush
app.opendash.cloud / northlake / tasksRead-only
Monday SaaS review
Monday SaaS reviewEvery Monday · 08:00 · Europe/Berlin
Scheduled
Slack#finance-weekly
OpenDash APP · 08:00

Your weekly SaaS review is ready.

Monthly recurring revenue$398.2kActive accounts · mrr v8

EMEA is down 6.1%. Brightwave and Kessler account for most of the change.

Uses the reviewed definition. The query and source rows travel with every figure.

Ask a follow-up, or correct me…
Sources ContextClaude Sonnet
0:30 / 0:30
Illustrative Northlake data

01 · TRACES

Every number
has a receipt.

Open any figure to see the query, definition and source rows behind it. A check written as code verifies the number against its recorded result.

Northlake · Inspect this answer
RECORDED RESULT

$398.2k, with the working.

SQL & sourcesDefinition
SELECT a.region, SUM(s.mrr) AS mrr
FROM subscriptions s
JOIN accounts a ON a.id = s.account_id
WHERE a.status = 'active'
GROUP BY a.region;
metrics/mrr · workspace definitionQuery checked
Sample source rows
AccountStatusMRR
Halden Roboticsactive$4,200
Pine & Foldactive$960
Brightwaveactive$4,100

Read-only query. Each figure points to its recorded result.

Illustrative Northlake data
Northlake · Improve this answer
“Past-due accounts aren’t paying customers.”

Saved for this conversation. Team knowledge goes through review.

PREVIOUS · mrr v7$412.6k
PROPOSED · mrr v8$398.2k
Curator proposes a precise changev7 → v8
− status IN ('active', 'past_due')+ status = 'active'

37 past-due accounts removed · $14.4k

Query passed48/48 evalsApproved by Maya
v8 released to the teamv7 retained. Revert when the definition changes.
Illustrative Northlake data

02 · LEARNS

Correct it once.
Make the fix
team knowledge.

The analyst remembers your correction. A proposed change then passes query checks, evaluation and an admin’s review before the team uses it.

Previous versions stay available. Revert when the definition changes.

03 · BUILDS

A dashboard built around
your next decision.

Describe who it is for and what they need to decide. Review the plan, watch it take shape, then publish a checked result.

app.opendash.cloud / northlake / dashboardsRead-only
New dashboard

Launch readiness

Checkout v2 · building your draft

Draft

Launch blockers

Waiting for source

Checkout events

Waiting for source

Revenue

Waiting for source

Waiting deals

Waiting for source

Sources stay attached to every widget. Publish when you’re ready.

0:00 / 0:10
Illustrative Northlake data

04 · RUNS

Your Monday numbers.
Already in the channel.

Write the question once and choose a time. OpenDash runs it again, checks the results and delivers a brief to Slack or email.

The same reviewed definitions. Fresh source data. A trace behind every figure.

See the Monday SaaS review
Northlake · Monday SaaS review

Monday SaaS review

Enabled

A question worth asking every week.

ASKMRR, net retention and churned MRR for last week, by region. Flag anything that moved more than 5%.
WHENEvery Monday · 08:00Europe/Berlin
SEND TOSlack#finance-weeklyEmail digest also enabled

Recent runs

Monday, 12 OctoberBrief delivered · query and source rows attached
08:00
Monday, 5 OctoberBrief delivered · query and source rows attached
08:00
Monday, 28 SeptemberBrief delivered · query and source rows attached
08:00

Uses the team’s reviewed definitions on every run.

Illustrative Northlake data
Northlake · The analyst in Slack
Slack#finance-weekly
DL
Dana 09:04

@OpenDash what drove the EMEA change? Was it mostly Brightwave?

OpenDash APP · 09:04

Yes. Brightwave accounts for $4.1k of the drop, followed by Kessler at $2.0k.

Brightwave −$4.1kKessler −$2.0kOther changes −$0.2k

Same metric definition. Same read-only sources. The working is attached.

Reply in thread…
Illustrative Northlake data

05 · IN SLACK

The next question
belongs in the thread.

Mention @OpenDash where the conversation is happening. It answers with the same definitions, the breakdown and a link to the working.

Your team can follow up without starting the analysis over.

How OpenDash works in Slack

Fits the way you work

Works with your stack. Any model you trust.

Connect your sources, choose your model, and get answers where your team works.

Your data

PostgreSQL
MySQL
ClickHouse
Snowflake
Databricks
BigQuery
Redshift

Jira, HubSpot and more through MCP

Your model · your keys

OpenDash

Read-only. Every number traced.

Where answers go

Slack

Answers with evidence

Email

Briefs and findings

Microsoft Teams Coming soon

Jira, HubSpot and more through MCP

Explore connections
Hosted in the EU
Made in Germany

Single sign-on OIDCSAML 2.0

A FEW DETAILS

Questions,
answered plainly.

How does it connect to my database?
Through a read-only user on PostgreSQL, MySQL, ClickHouse, Snowflake, Databricks, BigQuery or Redshift. It never writes to your database.
How is it different from a chat model?
OpenDash runs queries on your database. Every figure points to a recorded result, with its query and source rows available to inspect.
How are answers checked?
Code verifies each figure against the recorded query result. Corrections are checked, tested and reviewed before the team uses them, and can be reverted.
What does setup involve?
Connect a read-only database user and choose a model. Bring your own provider key or book a demo to walk through setup with your data team.
What recurring work can it own?
Routines run on a schedule and post to Slack or email with the trace attached. A watch can follow a metric. Your source data stays read-only.

Ask a real question.
See what comes back.

Bring a real question. We’ll show you the working.

Book a demo
Pair, the OpenDash mascot: a lavender plush