Marginal – See which customers and features drive your AI API costs
Demo account (shared, please be nice): https://marginalhq.com/login - email [email protected], password marginal-demo-2026. It's a seeded workspace for a fictional company; poke around Overview/Insights/Explorer. Or sign up and send real events.
What it does
Track every LLM call's cost and slice it by customer, feature, model — any field you define.
Marginal tracks the cost of every LLM call and slices it by the fields you care about — customer, feature, model. marginal.track({ provider : "openai" , model : response.model, usage : response.usage, fields : { customer : "acme-corp" , feature : "summarize" , }, }); One call per LLM request. Cost is computed server-side — usage in, dollars out. Slice spend by what matters in your product, see every dimension charted automatically, and check exactly what your integration sent. Insights charts spend for every dimension automatically — each model, provider, and registered field gets its own chart. No setup, no query builder. Every API request is logged with its outcome — accepted counts,…from marginalhq.com
Does the same job
all alternatives →- IBI built a service to help companies save 35%+ on their AWS bills2020 · ▲25
Hi everyone, I've been working on a side project over quarantine called Usage.ai and I finally feel comfortable enough to launch it. We're a service that plugs directly into AWS, automatically finds savings, and applies those savings at the press of a confirmation button all without ever needing to go to an AWS console. I'd love to get HN's thoughts on it! Demo: https://www.loom.com/share/2a6f1c8e4c214914a1cdd88c6fdec4ac Link: https://www.usage.ai/
- AAAdmin AI – follow up emails, notes and CRM updates from recorded call2023 · useadminai.com · ▲8
Hi HN! For the last 5 years I've been in sales (at mongoDB). Spent half my week writing updates for sales-managers. My founder told me this is a lot like JIRA. Updating tickets for internal management - I hear it's painful. Sellers call this "admin work". A lot of it is zero value (because it doesn't help us get paid), and easily takes up 5+ hours a week. Admin AI uses the recorded calls to draft follow up emails, write call notes and CRM deal updates. We're trained our AI to think like a seller. And also built all the integrations to make it feel magical. Love your feedback, comments,…
- FYFill your PDF templates with an API call - doqs.dev2022 · ▲6
Now live! I appreciate any feedback! If you think of subscribing, contact as at [email protected] and we will hit you up with a discount code!
- DTDashboard to Get Complete Breakdown of OpenAI Costs2023 · puddl.io · ▲5
Hello Folks, I am an indie developer, For the last 5 months I was freelancing by building GPT-powered chatbots. One of the challenges I faced was not having visibility into my OpenAI costs. This is why I have built a dashboard to get a complete breakdown of my OpenAI API usage through network calls from OpenAI. Please give it a try & help with feedback on how I can further improve the product,which can really add value and save some money. Note: 1)I found this API through the network logs of the billing screen. 2)The tool stores the OpenAI Key locally on the browser.
- IBI built an interactive demo builder for your sales team2024 · demoshake.com · ▲8
Hey HN! I’ve created (co-created with my co-founder) a tool called demoshake that helps you build interactive demos and use them in your sales process. The idea is simple - you create as many demos as you want, group them and let your sales team use them in their demos calls. You can embed demos in your landing page (iframe style) too. I welcome any criticism and suggestions! Harun
- LRLLM.Report – A free OpenAI API analytics dashboard2023 · llm.report · ▲5
Hi! My OpenAI costs started going up recently and I hated how I couldn't figure out which model was costing me more? The OpenAI usage page (https://platform.openai.com/account/usage) gives a nice overall breakdown, but it doesn't go into detail about cost / model, token usage, and # requests. So I reverse engineered their API to make a better dashboard, and then made it free to use. Try it out and let me know what you think. Also please tell me what other features you wish existed here https://platform.openai.com/account/usage but doesn't yet.
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, August 2026
the whole month →- TL
Life & fun · 10d ago · louisabraham.github.io
- SA
Hello HN! I found that picking out plausible but diverse skin tones for my digital art and game development projects was kind of difficult, and I got curious about if there was a way to define a color space that made it easy. I've built a color picker and procedural generation algorithm based on the space as well as a bunch of other fun js features and demos throughout the page that use the equations. If you find it interesting, I have lots of explanations of how I built it and what properties the space has. The methodology might be a bit shaky, but hopefully the result is as helpful for…
Life & fun · Aug 2026 · toneyalexander.github.io
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