Ideas Lab
Real problems people complain about online, pulled every morning and scored out of 100. Build, validate, or skip. How scoring works
- Ideas
- 4,820
- BUILD
- 702
- Sources
- 8
- Last sync
- 4h ago
Llmspeed – feel how fast a local LLM runs before you buy the hardware
Hacker News14d agoToolAI
Stoke Money – Personal finance app built around a cash flow forecast
Hi HN,I've been through many finance tracking apps, starting with Personal Capital, then Monarch, and most recently Copilot. None of them have worked perfectly for my needs.All of these apps tend to model “cash flow” as the net sum of your transactions during a month, yet while credit card charges are reported immediately, they are typically deducted from checking (cash) accounts the following month when the bill is due. As a result, I've always had to maintain a separate spreadsheet to track my actual cash flow.These apps also have a huge focus on managing budgets, which has never been for me. My spending is rarely consistent month-to-month, so every month I'm comparing my fully justified spending against my budget, and tweaking my budget until I “pass”.So I built yet another finance tracking app. It has the usual components - accounts, transactions, categories - but the core is a cash flow projection engine, and everything else is built around it.Budgets are gone. Instead, each income/spending category is projected forward from your historical activity, but not as a flat monthly number. They land in whatever account you historically use (e.g. checking account for paychecks/rent/bills, credit/debit cards for daily spending). Credit card payments are modeled automatically based on your projected credit card balance. You can set up your own recurring transfers between accounts as well, to model savings, for example.The end result is that the app can forecast how your cash moves, and you can see in advance when any account might run low. If all of your accounts look good for the next 12 months, maybe you don't have a budgeting problem.What’s made the biggest difference for me has been the MCP integration, which I have connected to Claude (but can be connected to any AI agent of your choice). I have asked it to import historical transactions by uploading PDF account statements, categorize/tag all transactions that took place
Hacker News14d agoToolAI
Things to Do API
I built an api/mcp for local events: pop-ups, one-off events, concerts, etc. currently covering the SF bay area.I've been building a bunch of consumer apps lately around personalization and recommendation and have had a hard time finding this data in one spot. agents can find useful stuff through tool calls / web search, but slow.you can scrape on demand too, but i wanted a faster data set and couldn't find an existing provider for the mix of event types I was searching for.the methodology for it is to find influential accounts/posters/etc. and do a fan out across things that they co-post with.mcp: `claude mcp add --transport http thingstodo https://api.thingstodoapi.com/mcp`I'm considering adding other cities. If you're building something that needs local event data, would something like this be useful and where would you want coverage?
Hacker News14d agoToolAI
Affordable AI problem-solving assistant for developers and businesses
There is demand for AI models like Gemini 4 Argon to solve complex problems, but access is often behind paywalls. A solo dev could build a cheaper AI-driven assistant to help developers and small businesses solve technical and operational problems affordably. [Trending: "gemini ai" with 100+ searches in GB]
Google Trends14d agoToolAIGB
rStation – A shared research lab for your AI agents
Hacker News14d agoToolAI
Ito – Build Terminal UIs in Lua
My friend and I are huge Neovim nerds. We like to rice our terminals and our editors. We just wanted a harness that we can do the same with, rice it till our hearts' desire. So, we started looking out for harnesses. Pi was great, Pi was good. But in the end, it was JavaScript. I don't have anything against JavaScript, but Lua is more fitting, and I can just as easily base it in Rust to make it even faster. So, we started cooking, and we ended up with a design that's minimal at its core and is built using Lua. Every built-in abstraction, the way you define providers, the way you rice it up, everything, literally everything, you can do in Lua.But how? How does it all work? I proposed we just do it the Ratatui way, but my friend was like, nah! We shall make it possible to build TUIs in Lua. (Weirdo!) But it paid off, and we got ourselves a shiny new Lua library, ito!!!He didn't get the gravity of the situation. We had an awesome terminal UI library in Lua! I wanted everyone to know how cool that is, so I got cooking, wrote myself a custom canvas implementation of this library instead of the usual terminal adaptor, exported it as wasm, and made this super cool website where you can experiment building awesome UIs in Lua. For vibe-coders, we got an llms.txt too. Have at it, and see if you can build something really cool with it. The library's at github.com/uji-labs/ito, and the awesome sauce harness is at github.com/uji-labs/uji.I hope you like it!
Hacker News16d agoToolAI
Indigene – Native plants for where you stand
Back when I lived in Central Florida, it took me years to learn what native plants could benefit wildlife and the local ecosystem as I slowly replaced parts of my yard to make it more resilient to drought. After moving to Seattle a few years ago, I had to start all of that from scratch. LLMs made the process a bit easier but I still had to routinely ask the same kinds of questions over and over, despite being far more familiar with native plants these days and leaning more on native plant societies and botany resources like Doug Tallamy and Joey Santore whose excellent Crime Pays But Botany Doesn't site was recently posted here[1].A good friend was visiting from Orlando over the summer and told me she was struggling to decide what to plant in her own yard despite her landlord being ok with it. That was enough of an excuse for me to kick off work on Indigene, which is aimed at giving folks just-in-time knowledge about what to plant based on their locale, sun exposure, soil conditions, and their own ecosystem support goals.Candidly, as many of us have, I've also leaned heavily on LLMs (mostly Claude here) to bring this to life. Largely to avoid saying demonstrably false things, which might be an absurd statement. You can find the GitHub repository[2] a wild array of open data sources I've dug out and cross-referenced to ensure mistakes and hallucinations are unlikely to survive for long. But given today's Indigenous Peoples Day I figured it was past time I shared the projects and invited more scrutiny.Being the Keep a Changelog guy, you can bet it has extensive release notes[3] generated from the project changelog. Those are also parsed into feature-specific guides so each also lists relevant recent updates.Lastly, I put together a post[5] that goes into a bit more detail on the story of the project and what I've learned along the way.I really, really hope people here (and beyond) find it as useful and enlightening as I do.PS: before it comes
Hacker News18d agoToolAI
My Terminal Themes Web App
Well, after 15 years of saying "Someday I'll turn this into a web app," I finally did it!I've been maintaining a popular repo of terminal themes at https://github.com/mbadolato/iTerm2-Color-Schemes, and it was finally time to build the app with all the features I wanted!So, the next-gen version is live at https://terminalthemes.com! It currently has over 800 themes, and provides downloads for close to 40 different formats (iTerm, Ghostty, Windows, and many more)It's free for everyone. An account is not needed to browse and download, but registering for an account (free) allows you to vote, bookmark, submit themes, etc.I focused on a11y features for those who need it, in addition to adding Find My Theme and other tools to make it easy to find something you like!Live terminal and code previews help show a more realistic view of the terminal, as opposed to the ANSI screenshots used at the GitHub repository. Those are still there, for those who like them.I would love for you guys to check it out and provide feedback, feature suggestions, whatever!And a huge thanks to the community of people that have contributed themes, ports to other terminals, and tools to provide conversions between formats over the years.Oh! And if you've ever contributed themes to the GitHub repo, please register for an account, find your theme(s), and hit the Claim Authorship button!
Hacker News16d agoToolAI
CAD for wood working with LLM support?
I’m a hobbyist woodworker. I manually design simple things in Shapr3d with a lot of patience when I build things. I wonder in the time of LLMs if there is an effective way to accelerate the 3D model design using AI.. Anybody tried using AI to model wood projects? Any tools to recommend?
Hacker News18d agoToolAI
I built a tool to track what AI leaders say and how their views change
Hacker News17d agoToolAI
I wrote an email app that accepts only one sender per inbox
Infinite inboxes, including prefixes where you don't have to create inboxes before you use them -- they self create when you add a suffix.Three types of inboxes: - One sender. Locks to the first domain to send email (except gmail, outlook, etc where it locks to an email account for obivous reasons. It bounces everyone else. - Temporary. Lasts for a while (24h, 7days, etc) then bounces. Received messages stay or go as per your choice. - Prefix. The most powerful. Declare a prefix like "service" and then invent any suffix, like "service-netflix" or "service$bank" and it creates a new one-sender inbox locked to whoever wrote first.Bring you own domain. Bring your own AI (MCP server ready). Forwarding.Please help me test it.
Hacker News16d agoToolAI
Petal, open-source CoScreen/ScreenHero finally free
Hi HN!I’m Jason Thomas, one of the founders of CoScreen.As some of you may know, Datadog sunset the CoScreen service on July 31st.Four of us from the original CoScreen team have spent the last couple months building a free and open source pair programming tool inspired by those that came before us: It’s called Petal (https://github.com/kiruna-labs/petal)Some history first, because when we started CoScreen we had hoped to break the sad and established pattern of this industry.Screenhero was an exceptional tool that a lot of us loved for pairing. Multiple cursors, low latency, you could click your partner’s editor, and just start typing. It felt like magic while we had it. Then Slack bought it in 2015, and shut down the standalone app in December 2017.When Till, Max, and I started CoScreen, we were aware of this history, and wanted to build an even more magical collaboration and screen sharing experience: the entire concept was “throw a window from your laptop up on someone else’s screen, and start collaborating with zero friction.”Datadog acquired us in 2022, and thus began the great adventure of adapting our real-time technology into Datadog products, while also continue developing CoScreen as a standalone tool from the Datadog ecosystem. Datadog invested in the product for years and we’re thankful for that. But CoScreen itself proved, like so many other screen sharing tool acquisitions, to be too tangential from the customer base and core mission. We instead built other communication features in Datadog, the integrations with other video platforms, and also its wildly successful MCP server. We weren't the only ones. Around was bought by Miro and shut down in March 2025. Multi (formerly Remotion) was acqui-hired by OpenAI in 2024 and shut down about a month later. It's nearly always the same story: a small team builds something hard that people love, a platform buys it for the team or the feature, and the standalone product doesn't
Hacker News16d agoToolAI
Pulse Local, a Mac budgeting app whose AI models run on your machine
Hi HN, I'm Niz. Pulse Local is a budgeting app for Mac (Apple Silicon, beta) that answers questions about your spending with models that run on your machine. Local mode needs no account. The ledger and your questions stay on the machine, and the app sends no analytics or crash logs.I felt there are not enough consumer apps that run locally with no calls to the outside world that help us manage sensitive corners of our lives. This app in its local variant only downloads a couple of models and occasional updates, it does not send any analytics nor data.Features-wise, Pulse local is a complete budgeting app that offers an assistant built on top of a fine-tuned model and thousands of rules embedded in the code itself. I'm sharing it for free with options to help me continue experimenting with fine-tuning and expanding along this direction.How it works: you import a CSV from any bank, in any currency. When you ask a question, a tuned Qwen3 0.6B model reads it into a structured query, and code, not the model, does the money math against the ledger. A 4B checker model double-checks each money answer; when the two readings disagree, the app says it is not sure instead of guessing. That change took wrong answers from about one in five to 5 in 494 questions across our last three tests. The write-up covers how we test, the designs that failed, and the limits of the claim: https://pulseos.finance/pulse-local/writing/agree-or-not-sur...The tuned weights are on Hugging Face (PulseOS/pulse-local-qwen3-0.6b).Limitations: the first start downloads about 2.9 GB of models. With the checker on, a money answer takes 12 to 14 seconds (you can turn the checker off). Small models can still misread an open-ended question. The app answers in English. Mac only for now; Windows is next.The app is free. An optional Connected mode (Sync Pass, $29/year) syncs US banks through Plaid; transactions pass through our server's memory and are not stored.
Hacker News17d agoToolAI
Parsimony – an open-source benchmark for how much code AI models add
Existing benchmarks measure solve rate and token output/cost, but don't punish excess code, which makes a codebase harder to work with in the long term. I wanted to measure this since it matters most for long-term projects.Method: it takes the patches that benchmarks already publish (DeepSWE, SWE-bench Verified, SWE-bench Live) and counts "coding units": syntax elements like statements, calls, names and literals. I didn't use lines of code, so cramming code isn't rewarded. Deleted code is subtracted, so it measures net units.GPT-6 and Claude 5.5 have no public patches yet, so I ran them myself on 3–10 tasks on my own subscriptions (shown as faded circles, not ranked).Any feedback or model contributions are welcome, especially public patch exports for newer models.GitHub: https://github.com/La5u/ParsimonyNot affiliated with any of the benchmarks.
Hacker News17d agoToolAI
Spotatui, a Spotify TUI/GUI without web API rate limits
About 11 months ago i revived the amazing but unfortunately unmaintained spotify-tui. Since then I have been working on it almost non stop, finding a lot of problems on the way.One of the main issues I ran into were the Spotify web API rate limits. Third party clients like spotatui are dependent on a developer app made in the Spotify developer dashboard, which get rate limited a lot and Spotify has slowly been deprecating endpoints from it. The workaround for this used to be using a shared client id from ncspot but unfortunately that got rate limited too. This is when i started adding other playback sources, including YouTube, Navidrome/Subsonic, Internet radio with +- 30.000 stations, local files, Qobuz for lossless audio and Apple Music for macOS only where it controls the music app remotely.I fixed both the deprecated endpoints and the rate limits with one solution. It still uses librespot in the back, my spotatui specific fork, but gets all data from it instead of only playback. The data comes from the internal API of the official clients, that API gives metadata, playlists, the user library, devices, lyrics and audio analysis. For searching it uses Pathfinder, the internal GraphQL API.There are no web API rate limits because there is no shared app quota. Every user uses their own session as if it was the official client. The only potential problem is that Spotify does not document this API so any change on their side would break it. I intend to fix any of those issues as soon as they pop up.It all started as a TUI back in 2019 but since then I've added a GUI to it too. The GUI is still a bit behind the TUI in terms of features, it only ships Spotify for now, but full parity between the TUI and GUI is coming soon. After that full parity with the official client is next. That is also why I made the GUI looks almost the same as the official client. That way its easier for me and users to find missing features.Spotatui also has lua scripting support, a op
Hacker News17d agoToolAI
Use your OpenAI subscription to create audiobooks
Hacker News18d agoToolAI
Consus (YC F24) – LLM gateway for companies handling sensitive data
Hey Hacker News! I'm Eric, the founder of Consus. We route all AI requests to ITAR restricted and controlled unclassified information (CUI) compliant infrastructure. Consus itself satisfies ITAR restrictions, and we recently completed our FedRAMP equivalency audit against the 410 FedRAMP High controls. FedRAMP High is the U.S. government’s strictest security baseline for cloud services handling sensitive unclassified data. A breach of this data could cause severe or catastrophic harm. Achieving FedRAMP equivalency, or authorization, is required before processing CUI information.Demo: https://youtu.be/6MYxJe6P4oYAerospace and defense is experiencing rapid growth. Unfortunately, most companies are struggling to use AI because of necessary data protections & compliance risk. These companies can move even faster if they could use the AI tools that most other industries take for granted.I spent most of my career prior to Consus in aerospace & defense. I launched national security payloads at SpaceX, and was an early employee at Castelion. I’ve watched first hand how compliance red-tape makes it difficult for companies to adopt new technologies and trust vendors that touch their sensitive data. For example, I worked with a satellite manufacturer that doesn’t allow any AI tools, even for non-sensitive work.Even as AI has become extremely useful over the past 9-12 months, I was surprised by how many aerospace & defense companies refused to touch it. They could see how it could be helpful, but they were scared about model providers training on their data, violating an ITAR rule, or risk losing new business because of a failed cyber assessment. These are all rational fears.There were a lot of ways to solve some/all of these problems, but we focused on building a compliant AI inference layer that can "plug" into any AI tool or workflow. Just as OpenRouter integrates with most tools, Consus can connect to tools like ChatGPT or Clau
Hacker News17d agoToolAI
AI agents unsuitable for irregular non-repetitive tasks
Very few daily tasks are repetitive enough to build automation for, and for irregular tasks it's hard to effectively delegate to AI agents.
X1mo agoToolAI
Improved AI video generator for specific fan content styles
Creators need better AI video tools that handle advanced camera cuts and avoid poor effects when generating Harry Potter or similar themed fan videos.
X17d agoToolAI
Agent37 (YC F26) – Low-cost agent sandboxes for running agents
Hey HN, I’m Vishnu from Agent37 (https://agent37.com). We build sandboxes for running AI agents at under 10% of the cost of most sandbox providers. Our customers spend an average of $0.37/month per sandbox.You can view a video walkthrough of the product here: https://youtu.be/DK5ISR2tbNoWe started as one of those OpenClaw wrapper companies back in February (yup.. judge me). In time we added more and more harnesses.One of the goals I had at the time was to offer a free plan to users, so as to gain more market share. But it was too cost prohibitive to do it right away. Especially since we were bootstrapped and servers were running on my savings.Today we’re actually able to offer the free plan thanks to the optimizations we’ve made since. Over 90% of the new signups we have are just free plan users and they host their personal Hermes/Openclaw/Claude Code etc on our cloud. Our infrastructure cost for the free sandboxes averages around $0.05 to $0.10 per user per month.Here are some techniques that have worked for lowering cost:* When you look at agent harnesses, they have certain patterns in the way they consume CPU and RAM. They need CPU in bursts when they’re initializing or performing certain tasks, but beyond that, they’re often not peaking. They need more RAM for things like browser and computer-use tasks. Once the task is done, usage often comes back down. This is a key insight in lowering cost, so by efficiently managing the CPU/RAM utilization we could lower the cost.* Secondly, you have to skip the hyperscalers when renting the Hosts. I worked at Amazon before and they have 80-90% margins on VMs. So it was an obvious choice once I looked at the prices to go one layer down to vendors like Hetzner and OVH. This alone was probably one of the biggest cost levers in reducing the cost.* Thirdly, you also want to pick a technology stack that lets you efficiently cache and share resources across the host. We use gVisor for b
Hacker News18d agoToolAI