Write JSON processor queries (jq) in natural language
Hey HN! Jaiqu is an AI agent for creating repeatable JSON transforms using jq query language syntax. Jaiqu translates complex JSON structures into repeatable jq queries. jq is a command line utility built for fast, reliable JSON queries. Jaiqu takes advantage of the LLMs for both constructing jq queries as well as using feedback to fix errors/invalid results if those queries are incorrect. Writing code to parse JSON is super annoying. It usually involves manually reading huge JSON balls, parsing the keys in a language of your choice, and pulling the data manually. Also, you're likely to…
In plain words
Jaiqu is an AI agent that converts natural language descriptions into jq queries for transforming JSON data. It helps users extract and restructure complex JSON without manually writing query syntax. The tool uses large language models to generate queries and iteratively fix errors based on feedback, addressing the tedium of manually parsing JSON structures and handling key mismatches across different data sources.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
Hey HN! Jaiqu is an AI agent for creating repeatable JSON transforms using jq query language syntax. Jaiqu translates complex JSON structures into repeatable jq queries. jq is a command line utility built for fast, reliable JSON queries. Jaiqu takes advantage of the LLMs for both constructing jq queries as well as using feedback to fix errors/invalid results if those queries are incorrect. Writing code to parse JSON is super annoying. It usually involves manually reading huge JSON balls, parsing the keys in a language of your choice, and pulling the data manually. Also, you're likely to run into key mismatches (i.e. source data might have "customer_id" as a key but my table is "user_id"), so you would have to write the translation dictionaries by hand. This struck me as a task that was analogous to the many AI "natural language to SQL" editors out there. However, there were a lot of consistency issues where the LLMs would not produce functioning code. So I wrote an AI agent that does it for me. It only takes two args: 1. A blob of messy JSON containing data you want 2. A JSON schema formatted in the way you want Jaiqu generates a repeatable jq query that collects the data in the exact JSON schema you specify. This way, you can run this query as many times you want or save it as a config without having to prompt LLMs each and every time. This tool beats the standard "prompt ChatGPT for a jq query" because by and large, LLMs don't perform too well when writing DSLs like jq. Jaiqu essentially works by: 1. Fuzzy matching each key against your schema to make sure it's available to query 2. Writing individual jq queries to fetch and transform those keys 3. Compiling each individual query into a single complex jq query. 4. Feeding back any errors or data mismatches to a "self-healing" loop that adjusts the queries accordingly until it works. Would love to hear thoughts or feedback. This has saved our team a ton of time internally, so hopefully it can help others as well.
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 2024
the whole month →
- 3Y
Life & fun · 2024 · github.com
Microlaunch▲1,116Launch and get feedback on both the idea and product
Dev tools · 2024 · microlaunch.net
