Querypanel

New on QueryPanel: faster natural-language analytics. Our v2 pipeline used to spend most of its wait time on classification-shaped LLM calls (guardrail, intent, schema linking, reflection). Those now run through Jev typed judgments, with the LLM as fallback. On our benchmark, first answers are ~54% faster and end-to-end demos ~44% faster. Details and numbers in the post. https://lnkd.in/eMQRipqe

Querypanel uses GPT 6 Solona and Luna now!

You can use Querypanel for internal dashboarding from now!

Schema sync tells the model which tables exist. It does not tell it what “active,” “risk,” or “revenue” mean in your product. We train QueryPanel four ways: - Gold SQL. Queries you already trust. The model copies that shape. - Annotations. Business meaning on tables and columns, not only usr_stat_cd. - Glossary. The words customers actually type. - Tenant definitions. Isolation field and per-tenant context, so “show all revenue” is still this customer. Start with five real questions. Fix the context. Expand after the answers are boringly correct. That is the whole semantic layer. Not a 40-page LookML project.

QueryPanel docs are now on Context7. If you use Cursor, Claude Code, or another MCP client, your agent can pull our current docs instead of guessing SDK APIs. That makes the first integration cleaner: attach a database, mint a tenant JWT, embed the dashboard, with today’s examples, not last year’s training data. Library: https://lnkd.in/dxxTNYxt Ask your agent about QueryPanel and add “use context7”. #embeddedanalytics #Context7

We added a full AI onboarding flow, you can use our MCP to set up Querypanel quickly. It - reads the workspace ID - creates the keys needed for JWT creation - writes the corresponding code for you Works for both the embedded Notion-like editor and the headless version.

MCP is table stakes for AI-native SaaS. So here's ours. Use QueryPanel from Cursor, Codex, Claude, or whatever LLM you live in: • create dashboards • add datasources • sync schema • train the system • or just ask questions and let the model analyze and write the report Happy querying. http://querypanel.io

We know data is a sensitive topic, so we shipped BYOK. 🔐 Some companies can’t use a shared OpenAI key. When restrictions are strict, they need generation to run on their own provider account, not ours. What it means in QueryPanel: 🔑 NL→SQL and chart generation run on your OpenAI key 🗄️ Keys are workspace-scoped and stored in Vault 👥 Owners and admins can rotate or revoke anytime Defaults we shipped with: ✅ Strict mode on by default ⚙️ Platform fallback is opt-in (auth/quota failures only) 🛡️ Moderation, embeddings, and evals stay on our key Embedded analytics shouldn’t force a tradeoff between product experience and data control. OpenAI BYOK is live in QueryPanel. 🚀

SEON is our first enterprise-grade customer, and this milestone means a lot to us. Working closely with their team on our headless SDK taught us a lot. It pushed us to support more use cases, think more carefully about flexibility and integration patterns, and improve the SDK in ways that make it a better fit for larger teams. The collaboration also surfaced edge cases that helped us improve the semantic layer behind our AI, which now produces better and more reliable results. This is one of the best parts of building: real customers force clarity. They expose edge cases, challenge assumptions, and help shape a stronger product than you would have built in isolation. Grateful for the trust, and excited for what comes next.

And here we go 🚀 , we got our first SMB customer here in Hungary. The company (https://lnkd.in/dGa-9XhU) produces sanitazer chambers that works with o3 for eggs 🐔 🥚 They are using our headless SDK to power their dashboards that helps them analyze chamber data. What an interesting company!