Tablize is the tool we wanted when we ran small teams with too much data and nobody to make sense of it. So we built it.
For the last twenty years, "talking to your data" has meant two things, and only two: hire a BI team to build dashboards (Tableau, Looker, Metabase), or hire someone who writes SQL (Jupyter, dbt, BigQuery console). For most teams, neither shape worked. They weren't analysts, and they couldn't justify a BI implementation for a question they'd ask once a month.
A Data Agent sits between those extremes. It's a chat surface backed by an agent that does the analyst work — picks the right tool for the question, writes the SQL or Python, runs the analysis, draws the chart, explains the result. Useful answers become reusable assets: reports, scripts, dashboards, watches.
We didn't invent this category. A handful of companies are taking serious swings at it. Our bet is on what happens *after* the chat — the keep loop is the differentiator, and we're betting the company on it.
Most "AI for data" tools assume you're technical. We assume you're not — and that the technical work (writing SQL, building dashboards, wiring schedules) is exactly the work you wanted the agent to do in the first place.
A chatbot answer disappears. A Data Agent answer becomes a Report, a Script, a Watch, a Dashboard — something that survives the session and works the next time you need it.
AI is great at confident-sounding answers and bad at being right about them. We built Verifiable Reasoning into the product so the agent slows down, samples your data, cross-checks itself, and shows you its math.
Self-hosting from Pro tier up. BYO LLM keys. Open-source runtime. The agent doesn't need to acquire access it doesn't need, and you should never have to choose between using the product and owning your data.
Primarily: small teams that don't have a data team. Solo developers and indie makers with product data but no analyst. DTC operators running Shopify with help from a friend. Freelance analysts setting up the same analysis on three different clients' datasets. IoT makers running sensors without a monitoring platform. Startup founders who want to start the week knowing what changed.
Not primarily: large companies with established BI departments. If you're already running Looker at scale with governed metrics and a data team that loves it, we're not the tool for you, and we'd genuinely encourage you to stay there. The tools you have are good for the shape you have.
Tablize is in active development. The core product — Connect, Ask, Keep — is shipped and usable today. Free tier is live, paid tiers run on Fly.io managed infrastructure, the runtime is open source, and the self-hosted path is well-traveled enough that we recommend it for any team with strict data residency requirements.
What we're working on next: deeper IoT, broader integration coverage, better forecasting primitives, more rigorous reasoning modes for high-stakes analytical work. The product roadmap moves with what our users actually use, which is more concrete than what they ask for.
Questions, ideas, complaints, partnership pitches, security reports — we read everything. Email hello@tablize.com. We respond to all real messages within a couple business days.
If you're a security researcher, please use security@tablize.com and see our security page.
The best way to understand Tablize is to use it for ten minutes on something you actually care about.
Try free with your data