frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Inconvo – open-source chat-with-data agent that doesn't generate SQL

https://github.com/inconvoai/inconvo
3•ogham•1h ago
Hey HN, we’re building Inconvo (https://inconvo.com) — an open-source tool to help developers build chat-with-data agents for their products.

Our open source repo is at https://github.com/inconvoai/inconvo (`npx inconvo@latest dev` to run locally), we also have a cloud version (https://app.inconvo.ai) :)

Here’s a quick demo of Inconvo: https://youtu.be/sbDTAbVG-WQ

We built Inconvo to solve a specific problem: building a chat-with-data demo is easy, but shipping to production for real customers is much harder.

We previously did a Launch HN (https://news.ycombinator.com/item?id=44984096). Since then, we’ve open-sourced the core: the entire data agent, semantic layer, and database connectors — under Apache 2.0, so developers can run it locally end-to-end, inspect the code, and contribute.

For a chat-with-data build, most teams start by showing an LLM the schema and letting it generate SQL. LLMs are great at SQL, so it's easy to get a demo working quickly, but going to production is where it breaks: you need to hide certain tables/columns, handle role based access to data and give the model a better understanding of your business logic and how it maps to the data. Some teams use prompts to attempt to solve these issues, but prompts are just suggestions to a probabilistic system that can emit arbitrary queries. For production, a set of guardrails and checks are required to make sure the system is behaving to your specification at all times.

Inconvo flips the approach: constrain the agent up front instead of trying to rein it in after. You define a semantic layer as the contract (approved tables/columns, metrics, filters/enums, join paths, tenancy rules). At runtime, the model outputs a structured query plan in the form of an IL (intermediate language) not SQL. We validate the IL deterministically against the semantic layer contract, and only then compile and execute it, so tenant scoping and field restrictions etc. are enforced by deterministic code on every request.

Here’s a simplified example of the structured query plan (IL) the model emits:

``` { "table": "reviews", "where": [{ "organisation_id": { "equals": 1 } }], "operation": "groupBy", "operationParameters": { "joins": [{ "table": "products", "joinPath": "reviews.product", "joinType": "inner" }], "groupBy": ["products.title"], "avg": { "columns": ["reviews.rating"] }, "orderBy": { "function": "avg", "column": "reviews.rating", "direction": "desc" }, "limit": 1 } } ```

In a nutshell, Inconvo data agents have a smaller set of safe moves they combine creatively, instead of an infinite surface area to validate after the fact.

There is a real tradeoff here: letting the model generate SQL is maximally flexible, and works well for systems with technical humans-in-the-loop who understand the data (e.g internal BI). Inconvo is more constrained by design so that behavior is enforceable for customer-facing use cases.

Because Inconvo is designed to be embedded in customer-facing products, we’re focused on the developer experience. Applications call Inconvo agents from their backend via our API or TypeScript SDK, and responses come back as typed, structured JSON (text, table, or chart as Vega-Lite spec) so they can be rendered directly in the UI. In multi-agent setups e.g. an application wide agent, it also fits nicely as a specialist chat-with-data sub-agent for the orchestrator.

We’re really happy to show this to you all. Thanks for reading. Please let us know your thoughts and questions in the comments.

Show HN: Hi.new – DMs for agents (open-source)

https://www.hi.new/
1•elieskilled•19s ago•0 comments

Ask HN: An approachable book on programming concepts/computer basics for a teen?

1•jamesgill•1m ago•0 comments

Malaysian Farmers Made a Fortune on Durian. Now It Is Piling Up

https://www.nytimes.com/2026/02/10/world/asia/durian-malaysia-china-demand.html
1•bookofjoe•1m ago•1 comments

'You can't just remove the cloud': US sprinting's reckoning before LA 2028

https://www.theguardian.com/sport/2026/jan/23/marvin-bracy-williams-doping-us-sprinting-paranoia
1•PaulHoule•2m ago•0 comments

Conductor removing support for Claude subscription authentication

https://twitter.com/charlieholtz/status/2024585923619590497
1•jrsj•3m ago•0 comments

A review of tumor-treating electric fields (TTFields): clinical advancements

https://pmc.ncbi.nlm.nih.gov/articles/PMC10476910/
1•CGMthrowaway•4m ago•0 comments

What's next for Chinese open-source AI

https://www.technologyreview.com/2026/02/12/1132811/whats-next-for-chinese-open-source-ai/
1•calcifer•8m ago•0 comments

Show HN: CursorLens – Open-source screen recorder/editor for product demos

https://github.com/blueberrycongee/CursorLens
1•blueberrycongee•8m ago•0 comments

Ask HN: What makes AI agent runtime logs defensible under adversarial audit?

1•catarina_eng•8m ago•0 comments

Show HN: Auto-scrolling for WhatsApp group chats

https://github.com/avremel/whatsapp-groupchats-extension
1•avremel•11m ago•0 comments

A Yale Professor's Investment Formula Says You Need More Stocks

https://www.wsj.com/finance/investing/yale-james-choi-portfolio-formula-stocks-02a96afb
1•sonabinu•11m ago•0 comments

A16Z has captured as much open source $ as all other early-stage VCs combined

https://twitter.com/PowersetRes/status/2024563164856123662
1•patrickdevivo•12m ago•0 comments

Show HN: Microstock.pro – AI-powered analytics for Adobe Stock contributors

https://www.microstock.pro
1•stokkerpro•12m ago•1 comments

There Is No Product

https://sidu.in/essays/after-ai-there-is-no-product.html
1•youknownothing•12m ago•1 comments

Luma AI is hiring SREs

https://jobs.gem.com/lumalabs-ai/am9icG9zdDqOhdaji2m34B3ibJdoy7Hb
1•SharLuma•12m ago•0 comments

Architectures of Error: A Philosophical Inquiry into Human and AI Code

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5265751
1•camilochs•14m ago•0 comments

U.S. Department of the Treasury's AI Strategy [pdf]

https://home.treasury.gov/system/files/136/Treasury-AI-Strategy.pdf
1•Nition•14m ago•0 comments

I thought no one cared about RSS. Turns out people do. Just not RSS itself

https://frido.app/apps/lume/
3•heymadsenx•14m ago•2 comments

Atlassian Founders Lose $7.2B as Software Stocks Slump on AI Fears

https://www.bloomberg.com/news/articles/2026-02-19/software-rout-wipes-7-2-billion-off-atlassian-...
2•speckx•14m ago•0 comments

Security Update Available for Metabase

https://www.metabase.com/blog/security-vulnerability
1•soheilpro•15m ago•0 comments

Highly Bespoke Software

https://twitter.com/karpathy/status/2024583544157458452
1•tosh•15m ago•0 comments

Show HN: Single Tool – Shorten links or send files instantly

https://www.singletool.io/
1•aliakyildiz•15m ago•0 comments

Show HN: I built Aegis AI – Agentic Home Security with GPT+Local VLM on Mac/PC

https://www.sharpai.org
1•simbaz•16m ago•1 comments

We hired and onboarded 3 new digital employees (OpenClaw bots)

https://www.veryfi.com/digital-employees-openclaw-bots/
2•ma-r-s•21m ago•0 comments

The End of Mail in Denmark – Who will mourn the last letter?

https://www.thedial.world/articles/news/mail-postal-service-denmark
1•TMWNN•21m ago•0 comments

People Who Won't Be Replaced by AI Are the Ones Who Outpace It

https://medium.com/@stefanvanegmond/the-people-who-wont-be-replaced-by-ai-are-the-ones-who-outpac...
2•stefanve•23m ago•2 comments

One Page of Async Rust

https://dotat.at/@/2026-02-16-async.html
1•birdculture•24m ago•0 comments

We're no longer attracting top talent: the brain drain killing American science

https://www.theguardian.com/us-news/2026/feb/19/trump-science-funding-cuts
12•mitchbob•24m ago•0 comments

Don't repeat yourself in metrics queries (2025)

https://scottnumamoto.com/dont-repeat-yourself-in-metrics-queries
3•scottnuma•24m ago•0 comments

Shafts – open-source SSH tunnel manager for macOS

https://github.com/xmstan/shafts
1•xmstan•25m ago•0 comments