Engraph – Automated ETL Pipelines
Hey HN, we’re Ross and Javier, co-founders of Engraph (www.engraph.ai). Our goal is to completely automate the process of building ETL pipelines, from ad hoc pipelines for question answering to fully fledged ETL pipelines within large organisations: For ad hoc pipelines, a question answering platform which enables users to ask questions in natural language about their organisation's data. Traditionally, access to data within organisations is limited to a handful of data-engineers. This means that if an employee needs access to some data, they have to go through a lengthy process of…
What it does
In the maker’s words, at launch
Hey HN, we’re Ross and Javier, co-founders of Engraph (www.engraph.ai). Our goal is to completely automate the process of building ETL pipelines, from ad hoc pipelines for question answering to fully fledged ETL pipelines within large organisations: For ad hoc pipelines, a question answering platform which enables users to ask questions in natural language about their organisation's data. Traditionally, access to data within organisations is limited to a handful of data-engineers. This means that if an employee needs access to some data, they have to go through a lengthy process of requesting it from a data-engineer, who then spends their day dealing with ad hoc requests. This process is time-consuming and inefficient for everyone involved. We solve this problem by passing natural language questions into a planner LLM that decomposes the task given the available data sources. This planner spawns workers that query the appropriate information from each individual data source, whether via SQL or searches across vector embeddings of unstructured information. Once the planner receives the data, it executes Python code to aggregate and operate on the data, and presents the answer to the user. For persistent ETL pipelines, instead of inferring the format of the output data from a natural language question users can provide a data output specification (e.g. YAML) through an API. This gets fed into a planner similar to the above. However, instead of running Python code to return an answer, we load the relevant data into an intermediate data lake. Internally (depending on the user’s preferences), we also use this API for the question answering platform: we learn from recurring patterns in the natural language questions (by clustering question embeddings), and implement persistent data storage that, in expectation, reduces the number of operations that need to be performed across the org’s data sources. For the user there is a trade-off between the cost of the intermediate data storage vs the cost/load of performing the same operations for every question. Surely you’ve been wondering about how we manage security and data access. Our goal is to never have to access company data on our side. Ideally all operations are performed on-prem. We take data privacy very seriously and we are building our platform accordingly. We charge on a per seat licence with lenient fair usage terms. We’d rather do this than charge per usage which we feel could disincentivize the use of the platform. If you're interested in trying out our platform, check out our demo video (https://youtu.be/Q8dNPQ8ofHk) or send us an email at {javier, ross}@engraph.ai. Thanks!
Does the same job
all alternatives →- PIParabola.io – Automate your work with visual programming2018 · parabola.io · ▲291
Hi HN! I’m Alex from Parabola (https://parabola.io). Parabola is a visual programming tool for creating functional data flows that everyone can use. It’s entirely drag-and-drop, handles data sizes much larger than a traditional spreadsheet, calculates everything live, and can run your flows on a schedule of your choosing. I used to work in strategy consulting, doing data analytics for SMBs and Fortune 500 companies. The amount of time wasted on menial tasks was astounding. Things like cleaning data, generating custom reports, creating human workflows to solve shortcomings in third…
- PLPipelines – Language for scripting parrallel pipelines with Python2019 · github.com · ▲69
- HAHyperFlow – A self-improving agent framework built on LangGraphApr 2026 · ▲8
Hi HN, I am Umer. I recently built an experimental framework called HyperFlow to explore the idea of self-improving AI agents. Usually, when an agent fails a task, we developers step in to manually tweak the prompt or adjust the code logic. I wanted to see if an agent could automate its own improvement loop. Built on LangChain and LangGraph, HyperFlow uses two agents: - A TaskAgent that solves the domain problem. - A MetaAgent that acts as the improver. The MetaAgent looks at the TaskAgent's evaluation logs, rewrites the underlying Python code, tools, and prompt files, and then tests the new…
- ACAgent-contracts, contract-based LangGraph agentsJan 2026 · github.com · ▲8
Hi HN, I’m the author of agent-contracts, a Python library that explores a contract-based approach to structuring LangGraph agents. When building larger LangGraph-based systems, I kept running into the same issues: - node responsibilities becoming implicit - state dependencies spreading across the graph - routing logic getting harder to reason about - refactoring feeling increasingly risky agent-contracts is an attempt to make these boundaries explicit. Each node declares a contract that describes: - which parts of the state it reads and writes - what external services it depends on - when…
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
Launched alongside, March 2023
the whole month →
- BI
I'm a big fan of the BBC podcast In Our Time -- and (like most people) I've been playing with the OpenAI APIs. In Our Time has almost 1,000 episodes on everything from Cleopatra to the evolution of teeth to plasma physics, all still available, so it's my starting point to learn about most topics. But it's not well organised. So here are the episodes sorted by library code. It's fun to explore. Web scraping is usually pretty tedious, but I found that I could send the minimised HTML to GPT-3 and get (almost) perfect JSON back: the prompt includes the Typescript definition. At the same time I…
AI · 2023 · genmon.github.io