Show HN: Onlyrecipe 2.0 – I added all features HN requested – 4 years later

📝 Discussion Summary (Click to expand)

The three most prevalent themes in the Hacker News discussion about OnlyRecipe are:

1. High User Demand for Accurate Unit Conversion and Recipe Scaling

Users strongly value the utility features, particularly the conversion between US Customary and Metric units, although they are highly critical of the initial accuracy. There is also significant interest in scaling recipes (adjusting servings).

Supporting Quotes:
- On accuracy issues: "The example of 2 Cups of greek yoghurt being 391.32 g... This leads me to calculate that 2 cups greek yoghurt weighs between 454 g and 499 g). The 391 g value is way off." (pjsg)
- On the precision needed for baking: "There is no (serious) baker that use volume for flour. If the recipe uses volume, it is flawed and you should find another recipe." (ansc)
- On scaling: "Something I'd like to see in a recipe app is scale is not only reflected in the ingredients list but also in the directions." (jpalmer)

2. Criticism of Traditional Recipe Blog UX and Content Bloat

A primary driver for using the tool is the user frustration with poorly formatted or distracting recipe websites, characterized by excessive personal stories and ads.

Supporting Quotes:
- On the core problem solved: "That initial momentum proved that the core problem (ads, life stories, and clutter on recipe blogs) needed a solution." (AwkwardPanda - describing the need)
- On anecdotal fluff: "It's not just 'taking other people's suffering and making it about oneself.' It's often purely SEO." (Telemakhos)
- On frustrating UX: "The fact that I couldn't scroll to the actual recipe was infuriating. You're despicable!" (port11)

3. Technical Implementation Hurdles and Required Polish

Users provided detailed technical feedback regarding platform-specific issues (Flutter/web rendering), data parsing reliability, and basic UI/UX necessities that needed immediate fixing.

Supporting Quotes:
- On text selection: "A major point is that I don't seem to be able to select text to copy and paste." (regularfry)
- On parsing failures: "Doesn't seem to work for this blog: https://foodwishes.blogspot.com/..." (xiconfjs)
- On performance/UI: "The animations are laggy and the transitions make the website slow." (bflesch)

🚀 Project Ideas

Recipe Import LLM Refiner & Standardizer

Summary

A service/tool that uses an LLM (like ChatGPT, as hinted by the developer) to aggressively clean, standardize, and structure recipes that fail standard HTML/JSON-LD parsing.
Core Value Proposition: Guaranteed recipe extraction from virtually any web source, specifically solving the issue of parsing obscure blogs (like FoodWishes) by translating messy text into a consistent, structured format usable by recipe management apps.

Details

Key	Value
Target Audience	Power users of recipe apps (like OnlyRecipe, Paprika users wanting to migrate), developers needing reliable recipe data feeds, and creators of clean recipe websites.
Core Feature	LLM-powered inference engine to ingest unstructured text and output standardized JSON adhering to a strict schema (including handling ingredient density/conversion inconsistencies).
Tech Stack	Python/FastAPI backend, OpenAI/Anthropic API for inference, PostgreSQL for storing standardized schemas, lightweight front-end validation tool.
Difficulty	Medium
Monetization	Hobby

Notes

Why HN commenters would love it: Directly addresses the failure points mentioned by users like xiconfjs ("Doesn't seem to work for this blog: https://foodwishes.blogspot.com/") and the developer's own plan to use an LLM as a fallback. It formalizes this crucial feature.
Potential for discussion or practical utility: Professionalizing LLM data extraction for niche datasets like recipes is a hot topic. Could be spun off as a B2B data cleaning pipeline for culinary startups.

Precision & Consistency Unit Conversion Validator (SigFigs & Context)

Summary

A specialized utility that refines the existing unit conversion feature by focusing on appropriate significant figures (SigFigs) and contextual unit selection (weight vs. volume).
Core Value Proposition: Eliminates user frustration over overly precise (e.g., "453.59 grams") or contextually incorrect conversions (e.g., converting liquid volume of yogurt to mass incorrectly), enhancing measurement utility, especially for baking.

Details

Key	Value
Target Audience	Home bakers, users transitioning between Imperial/Metric, and users who prefer rounder metric figures (`regularfry`, `pzo`).
Core Feature	Post-processing step on unit conversion: (1) Applies context-aware SigFig rules (e.g., 2-3 SigFigs for most kitchen measurements). (2) Offers the choice between volumetric (mL) or mass-based (grams) conversion for ambiguous ingredients (e.g., yogurt, flour), respecting user knowledge (`folli`, `cengen`).
Tech Stack	Frontend logic (Flutter) to manage user preference toggles, backend service utilizing a curated ingredient database mapping volume to density (weight) derived from reliable sources (like King Arthur reference data).
Difficulty	Medium
Monetization	Hobby

Notes

Why HN commenters would love it: Directly addresses the math concerns (pjsg questioning the 391g yogurt conversion) and the significant figure debate (regularfry, netsharc). It shows an understanding that "precision" isn't always "accuracy" in a home kitchen context.
Potential for discussion or practical utility: It touches on the recurring tension between baking precision (weight for flour) and general cooking convenience (volume for liquids), offering customizable paths.

Recipe Step Synchronization & Contextual Highlighting Tool

Summary

A tool that ensures recipe instructions are scaled correctly alongside ingredients and provides dynamic inline highlighting based on the current ingredient list.
Core Value Proposition: Solves the problem of instructions not reflecting scaled ingredient amounts (steps referring to "1/2 teaspoon" even when the recipe is doubled) and allows users to instantly find where an ingredient is used in the steps.

Details

Key	Value
Target Audience	Users who scale recipes up or down (`jpalmer`, `pzo`), and power-scrollers who need to follow along without losing their place (`sequoia`, `meowcow`).
Core Feature	(1) Step Scaling Logic: Automatically substitutes scaled ingredient quantities into corresponding steps if the original recipe failed to do so, with optional unit conversion applied to the instruction units as well (`Robbsen`). (2) Dynamic Highlighting: Renders ingredients and their corresponding scaled quantities highlighted (optional feature) in the instruction text, allowing instant cross-reference between the ingredient list and the narrative flow.
Tech Stack	Client-side JavaScript/Flutter logic to manage state between the ingredient list and instruction rendering. Requires robust mapping from the parsed ingredient dataset to textual references within the instructions (potentially using NLP or basic token matching).
Difficulty	High (due to the complexity of mapping textual references in free-form instructions to structured data fields, especially when instructions include measurement phrases like "a pinch of salt" vs. structured "1/4 tsp salt").
Monetization	Hobby

Notes

Why HN commenters would love it: Addresses the workflow issue where scaling is applied only to the manifest, not the instructions (jpalmer). The highlighting feature satisfies requests for better readability and reduced cognitive load while cooking (sequoia, meowcow, mdip).
Potential for discussion or practical utility: This feature tests the boundary between strict data parsing and understanding natural language instructions, a highly valuable but difficult feature for automated recipe tools.