Launched this week
Tabloy AI
Local-first desktop analytics workspace with AI assistance
10 followers
Local-first desktop analytics workspace with AI assistance
10 followers
Start for free and use Chat, Tabloy Agent, and local-first dashboards with CSV, Excel, JSON, and SQL data, also KoboToolbox data. Keep dataset processing local while AI works from schema context.
Tried Tabloy AI on a messy CSV and liked that nothing left my machine, even while the agent ran queries against it. The schema-aware chat caught column mismatches I usually miss in Excel.
@pakizeakcqewn Thanks, that’s exactly the workflow we designed for. Messy CSVs stay on your machine, and the agent only works with schema/context so it can catch column mismatches early. Super happy it caught things Excel usually hides. If you hit a weird schema case, feel free to share (no data needed) we’re always tightening that path.
Loving the local-first angle here, feels rare for AI tools. One thing I'd love to see is a scheduled query or dashboard refresh option, so the Tabloy Agent could automatically re-run analyses and update charts overnight. That would turn it from something I check into something that surfaces insights to me on its own.
@tunahanalaukfl Thanks Tunahan, really glad the local-first approach resonates. Scheduled refreshes / overnight agent runs are high on our list: same local runtime, just re-running saved analyses and updating the dashboard without you babysitting it. We’re shaping that next so Tabloy can surface insights instead of waiting for a check-in. Appreciate the push!
The local-first approach is genuinely useful, especially for working with survey data from KoboToolbox in the field. One thing I'd love to see is a way to share a dashboard read-only via a simple link or export to a static HTML report, so colleagues without the app installed can still review findings without needing a login.
@diyardikicjrbg Thanks so much for the feedback! We're thrilled to hear that the local-first approach is helping with your work—that’s exactly the kind of use case we built it for.
Sharing via a simple read-only link or exporting to a static HTML report is a fantastic idea! It definitely makes sharing insights with colleagues painless, without forcing them to install anything or sign up. We’ve added this right to Tabloy AI.
Stay tuned for updates, and thanks again for helping us make the product better!
@diyardikicjrbg In the Tabloy AI dashboard window, there is already a ‘Share’ button in the top-right-hand corner for creating a static HTML report.
A small thing that would help me a lot: add a way to schedule the Tabloy Agent to run a saved query on a cron-like schedule and email me the resulting dashboard snapshot. Right now I have to remember to rerun things, and the local-first setup feels perfect for this kind of background automation.
@elifnur1388706 Love this, Elifnur. You can already share a dashboard snapshot by email today, what’s missing is the cron-like “run this saved query / agent job and send me the result” loop. That’s a natural next step on top of local-first + email share, and your note is a clean product brief. We’ll keep it close while we ship scheduling. Thanks for spelling it out so clearly!
Loaded a CSV through the agent and it caught a schema mismatch I'd missed in another tool. The local-first angle is the real win for anyone handling survey data.
@nurcanc8nn Thanks Nurcan, schema mismatch catches are one of our favorite “quiet wins.” And yes: survey/sensitive datasets are a big reason we went local-first. Your raw responses never need to leave the machine for the agent to be useful. Glad it clicked for that use case!
Love how it keeps processing local and just pulls schema context for the AI, that split honestly feels really well thought out instead of just tossing everything into the cloud.
@brahawj Thank you Büşra! That split is intentional: heavy lifting stays local, and the model only gets the schema context it needs. We didn’t want “upload everything to the cloud” as the default for analytics. Means a lot that it feels thoughtfully built, not just another AI wrapper.