ZAANIX Agent

Hand it the question. It works as you, and asks before it changes anything.

ZAANIX Agent is built into the platform: Home is the agent. Choose the data and say what you need. The agent plans, runs the platform's tools with your permissions, and brings back findings, charts and the SQL behind them.

Run the agent Read the guide

A mission in ZAANIX Agent: the question, the answer with key findings, the query results as charts, and the canvas with pin, save and CSV actions

Missions, not chats

Each request is a mission you can come back to, continue, duplicate or archive.

A plan, then the workThe agent says what it will do, then does it one step at a time. The steps are folded away and one click from the full detail.

Answers with their evidenceKey findings, and every result as a chart, a table or its SQL — marked "verified metric" when a defined metric answered.

Chosen and foundThe data you chose and the data the agent found are shown apart, so you know what the answer rests on.

What to ask nextFollow-up suggestions after each answer; @ to mention data and / for common requests.

The canvasOne result shown large beside the thread: switch the chart, pick the axes, see where the numbers come from.

Live everywhereMissions update as they run, in the thread and in the sidebar, grouped as Needs you, Running and Today.

Context you choose

The data lives beside the prompt, not buried in it. Add it from anywhere you can read it.

The agent home: a prompt with Analyse data chosen, capability chips, sample prompts, and the Context panel

In the workspaceTables, views and files, as the catalog shows them to you.

Server filesFolders and data files on the ZAANIX server, as its file jail allows.

RemoteLakehouse tables (Databricks, Iceberg catalogs), database connections, cloud buckets, or any URL DuckDB can read.

From your computerDrop files on the panel; they are uploaded into the workspace and added.

CapabilitiesAnalyse data, build a dashboard, chart, check quality, investigate, clean data, notebooks, data apps and metrics — shown when your role allows them.

Sample promptsChoosing a capability shows requests written for the data in your context, and points the agent at the right tools first.

In control

The agent works with your access — and no more.

Sign-in is passed through to ZAANIX. Reads use your session; tools use an agent token minted for your sign-in, so ZAANIX applies your roles, row filters and column masks to every call, and audits it as the agent's.

What the agent does	Examples	By default
Read	Queries, profiles, the catalog, lineage	Runs
Create things	Dashboard widgets, quality checks, notebooks, saved queries	Goes ahead
Change or delete data, or export it	SQL that changes tables, models, writing files	Asks you
Publish or reach outside ZAANIX	Apps and endpoints, alerts, reverse syncs	Asks you
Drop, delete or truncate	Anything that removes data	Always asks

The autonomy dialEach person sets Ask me or Go ahead per kind of action. After several approvals in a row, it suggests letting the agent go ahead.
An approvals inboxEverything waiting for you, across missions, with the statement and the reason.
An activity logWhat the agent made and every approval — asked, given, declined or automatic.
Private by defaultMissions and briefs are yours. Sharing happens in ZAANIX, under its access rules.

Results you keep

Nothing stays trapped in the conversation.

A brief: the mission as a report page with a summary, key findings and a chart

Pin to a dashboardAny result, as a chart or a table, on a new or existing ZAANIX dashboard.

Save as a queryThe SQL saved in ZAANIX, or opened there in a query tab.

BriefsA mission as a report page — summary, KPI tiles, charts, tables and findings — that you edit, refresh and print.

Shared as a dashboardSave a brief as a ZAANIX dashboard; that is how it is shared, under ZAANIX's access rules.

Everything it builtDashboards, notebooks, apps and checks the agent made open in the platform.

A CSV when you need oneDownload any result, or copy a brief as Markdown.

Your model

Set one for everyone on the server, or let each person bring their own key.

ProvidersAnthropic, OpenAI, OpenRouter, Ollama and any OpenAI-compatible endpoint.

Open models understoodThe agent reads the native tool-call formats of open models (Qwen, GLM, Llama, LFM), not only its own.

Keys stay privateA person's own key is sealed on the server and used only for their requests, everywhere they ask.

Cost on recordTokens, latency and an estimated cost for every task, by model, from the telemetry API — at your own prices if you set them.

Built into ZAANIX.

Since 3.0 the agent runs inside the server: Home is the agent, every page has Ask about this, and missions live in the Library. Set a model and it is on; leave it out and ZAANIX works without it.

Deployment The guide

Available on request

ZAANIX Agent comes with ZAANIX. Talk to us to try it on your data, with your own model or the one you choose for your team.

Talk to us The guide

Talk to us about ZAANIX.

ZAANIX is available on request. Tell us about your data and your team, and we will set you up with access, the deployment files and a walkthrough.

Email us Message on LinkedIn

or write to ajmalnazirbaba@gmail.com