← Lab
tools·In progress·2026

Huella

A tool to plan a long trip, adjust the route without breaking the budget and keep the journey as I go.

Logbook · 11 entries ↓Ver demo ↗
01

The Idea

I want to travel to China in 2027. The trip is still being planned, I start alone in Beijing, and there are enough cities, attractions and pending decisions for a spreadsheet to start looking like a threat.

Huella was born to put all of that in order.

It is a progressive web app to plan and document the trip: see the route on a map, organize the stops, calculate nights and costs, and record by GPS the places I went through.

The budget is calculated per person. The exception is lodging, which is split among those who share a room. Because traveling with someone shouldn’t mean the budget stops making sense.

The idea is to plan with the comfort of a big screen and check the trip from my phone. I also want it to stay useful when there is no connection: an itinerary that disappears right when you need it isn’t an itinerary, it’s decoration.

The project started out called GoChina. It later became Huella (“footprint”): a name that doesn’t limit it to one country and that also speaks of the trail a trip leaves behind.

02

The Why

I’ve wanted to make this trip since I started studying Chinese.

So far I’ve known part of the country through the language and through the people I’ve met while learning. I want to take that further: meet them in person and see with my own eyes places I only know from a distance.

I’m interested in seeing the China that is today a technological power and letting what I find inspire me. But I also want to walk its villages and ancestral places. See those two extremes on the same trip and photograph them.

And to make a trip like that, you first have to make a lot of decisions.

Each one moves another: adding a city changes the transfers, moving a stop shifts the dates, and staying one more night affects the budget. If I also want to compare options without erasing the previous ones, things get interesting. And by “interesting” I mean I end up negotiating with seven tabs and a cell nobody dares touch anymore.

I needed to see the whole trip and be able to play with it: change the route, check how much it costs and try scenarios without starting over.

Huella is also a way of documenting how I make decisions while I build. Not only what works, but what still needs checking, what I had to change, and the numbers that don’t yet deserve to be called final data.

03

Logbook

11 entries
Process
Lesson

Changing computers without starting over

I cloned the repository on another computer and got the project running in a few minutes.

The difference wasn’t magic or an exceptional setup: I had written down the project’s state, the decisions and the steps needed, without including secrets.

Documenting while I build isn’t only for telling the process afterwards. It also keeps me from depending on remembering how I solved something three days ago.

Research
Decision

A useful link beats an invented “official” one

I wanted every attraction to have a link to look up information and see photos. I tried to solve it with official sites, but I couldn’t verify links to include.

I went with Wikipedia: I searched the attractions by their name in Chinese characters and picked the Spanish article when there was one; if not, the English or Chinese one.

According to the recorded check, 76 of the 78 attractions have a link and all 76 responded. Two have no link, and none of the official sites ended up verified.

It’s not the definitive source for prices or opening hours, but it’s a starting point. Better to leave a blank than to send someone to the wrong page with great confidence.

Product
Progress

What if I skip it? Now I can try

I wanted to compare routes without deleting cities every time I changed my mind.

Each stop has a checkbox to deactivate it. The city stays in the list but stops counting in the budget, the nights and the map. If I reactivate it, it adds up again as before.

It looks like a small interaction, but it changes the way I plan: I no longer have to commit to a decision in order to explore it.

Product
Insight

One night at every stop wasn’t an itinerary

Reviewing the route, all 37 stops had a single night assigned. It was the minimum possible, yes. It was also a rather efficient way to turn a trip into a suitcase-packing competition.

I added two references per city: a minimum to see the essentials and a recommended length to travel with less rush. In the recorded calculation, those references added up to 35 and 63 nights, respectively.

They are my own criteria based on typical guides, still without verified sources. They’re not meant to decide for me: they’re there so the length of the trip stops being a default number.

Interface
Progress

A route to share, not a file to decipher

Exporting a JSON can work for whoever manages the plan. For someone who just wants to know where we’re going, not so much.

I added “Copy route”: a text version of the itinerary, with stops, dates, transport and total per person.

The idea is simple: that I can send it to someone without having to explain what to do with a file full of braces and quotes.

Infrastructure
Milestone

Huella reached production. Permissions are still under review

The app is now live at huella.davila.uno and connected to Firebase.

In the recorded test, an attempt to read the plan without signing in got a 403 response. That confirms a block in that case, but it also leaves an open question: how will someone who only wants to look at the itinerary get to it?

Before calling read mode settled, I have to check the full flow and test the rules with an invited person.

Product
Decision

Sharing the route shouldn’t mean handing over the keys

The plan holds my dates, my costs and decisions I’ll keep adjusting. I want to be able to share it without just anyone being able to change it.

The intention is that other people can look at it and only I, or whoever I invite, can edit it. The control doesn’t rely only on disabling buttons: the data layer must also reject unauthorized changes.

Because a greyed-out button isn’t security. It’s a well-behaved suggestion.

Pending: clarify read access without signing in and test the permissions with an invited person.

Infrastructure
Pivot

Switching providers without emptying the wallet

I had picked Supabase to store the plan in the cloud. The problem was that the two active projects of the free plan were already taken by Rènshi.

I could pay or move something that already worked, but neither option made much sense. These projects are also my space to play, experiment and try technologies; I don’t need to turn every curiosity into a new subscription.

So I switched to Firebase, which I already knew from Nimbus. Huella still needed a way to store the plan, but not at the cost of affecting another project or spending more than necessary.

Experimenting is fine. Paying rent for every experiment, not so much.

Product
Decision

GoChina fell short

The first name described the destination well, but it locked the tool in there. If I later wanted to use it for another trip, GoChina would have some explaining to do.

I renamed it Huella. It works for different destinations and connects with another part of the project: keeping the route and documenting what happened during the trip.

I considered other names, like Lutu and Xingcheng, but Huella said what I needed without tying itself to one route.

Product
Doubt

A round-trip flight, a route that still doesn’t close

I added roughly S/ 8,000 to the budget for the Lima–Beijing flight, converted to yuan. The amount covers a round trip and assumes buying it well in advance. It is still an estimate, not a confirmed fare.

Another inconsistency also showed up: the route starts and ends in Hong Kong, although the intention is for it to start and end in Beijing.

That stays pending. I’d rather the log point it out now than find out when the itinerary already looks finished.

Product
Decision

A budget can start with doubts, but it has to say so

The first version gathers 78 paid attractions in 31 cities. Of those, 38 are marked as must-sees and add up to roughly ¥2,518 per person.

The prices come from an initial search done with an AI agent. They work as a starting point but still need checking, so they show up as “est.”. When I edit one, it is labelled “yours”.

The app can help me calculate with what I have today, but that doesn’t turn an estimate into a confirmed price. I wish budgets worked like that.

04

Current status

In progress
Ver demo ↗