Launched this week

Goldcrest
Your financial plan, one question at a time
2 followers
Your financial plan, one question at a time
2 followers
I built Goldcrest in Toronto for my own use in the hope it helps others. I was discussing financial planning (specifically a rent vs buy scenario with my wife) and found that it requires a surprising number of variables and complexity to do proper modelling/forecasting (the model implementation is a Monte-Carlo simulation under the hood). The idea is an AI chat agent reduces the complexity and tediousness of form-filling, and will help prompt users similarly to a financial advisor.



Free Options
Launch Team / Built With

AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted



I built Goldcrest in Toronto for my own use in the hope it helps others. I was discussing a rent vs buy scenario with my partner and found that it requires a surprising number of variables and complexity to do proper modelling/forecasting (the model implementation is a Monte-Carlo simulation under the hood). I then extended it to financial planning since most of the groundwork was already built from the rent vs buy model.
I did look at some other financial planning tools and noticed that the main difference between all other tools and my workflow was that I was primarily interfacing with my LLM to discuss, learn, plan, and build my models vs filling out form inputs in existing apps (rather tediously).
So, I thought that could be something that other people would like to do as well. Most of my friends aren't in software, and if they use LLMs it's for fairly basic things. For me at least, it's easier to type into a chat, "Hey, I make $x/year, have a wife and kids, a house valued at $x, a TFSA/Roth IRA with $x, etc." Right now I have the chat agent prompting the user. It's a similar experience to how one might have a conversation with a financial advisor, however with slightly more guardrails (e.g. the LLM never does math, just translates your input to form inputs).
The other tedious aspect of financial planning that I unfortunately can't solve right now is financial data integrations (investment platforms, banks, etc). US banks and institutions have more support for integrations so this is on my radar, but Canadian institutions are quite behind so this will take some time. So users have to collect their financials on their own for now.
I also wanted to put an emphasis on privacy given my personal detest for companies that don't value it, so everything is pseudonymized (similar to anonymization). Your data should be yours.
As for the model, it leverages open-source data to ensure everyone (well, USA and Canada at the moment) has a tailored forecast that calculates in-depth tax considerations for various registered investment accounts, full-time vs self-employed income, detailed future expense considerations and "events" like having a child or selling a house, Social Security/CPP/OAS, etc. Everything is also manually reviewable and configurable for users inclined to do so. I'm working on an educational aspect/feature so users can ask the LLM questions. I also have plans to expose an MCP for power users so you can hook-in with your own LLM.
There's also a Sankey diagram for income breakdown. Everyone loves a Sankey diagram.
Thanks for reading.
I'd love to hear your thoughts and feedback.