Fig does not send sensitive data to the cloud. Our products (autocomplete, sidebar, apps etc) process data locally on your device.
Fig tracks basic usage metrics (outlined below). This enables us to make better product decisions and provide better support. During our private beta, this telemtry will be de-anonymized. After publicly launching (ie Fig is accessible for anyone to download), all telemtry will be anonymous.
If you ever have any questions about how Fig manages your data, please feel free to email us at any time: firstname.lastname@example.org
What Information Fig Collects and WhyCLI Usage
Each time you run a fig cli command (like
fig update or
fig docs) we send a ping. The CLI is an important part of the Fig app. If
fig update break, up to date completion specs won't be downloaded for users,
meaning Fig will break.
We send a ping each time you select an autocomplete suggestion. The ping contains the root command you used (e.g. git, cd, aws, etc) and nothing else (ie if you select "commit" in "git commit", all we see is git). As we have a small userbase, this currently occurs for each suggestion you select. As we scale, this will not be necessary. We will then phase it out in favour of the Terminal Usage ping below.Terminal Usage
Sometimes developers use the Terminal 100+ times a day. Other times they don't use it at all. If somone uses Fig 50 times one day but 0 times the next day, is it because they turned off Fig or because they didn't use a Terminal? Knowing this without being instrusive will help us make better product decisions.
Therefore, once a day, Fig sends a single ping with aggregated metrics on how long you used the Terminal, how many times Fig popped up, and how many times you used Fig.App Usage
Fig sends a ping for basic app events like login, quit, and turning autocomplete on/off.
Information Fig Does Not Collect
Other than what is mentioned above, Fig never sends data from your Terminal or device to its servers unless you specifically opt into it.
Why is your tracking de-anonymised during your private beta?
Our aim with de-anonymising telemetry during our private beta is to work more closely with our users. When we say closely, we actually want to personally get to know our first 1,000 users. We see de-anonymity as a two way street: if you are willing to try our product and let us know who you are, you should 1. know who we are and 2. have a say in shaping our product and vision.
Fig is still early. Making product decisions in the early days with minimal data is hard. To get this data we could track everything and never speak to you. Alternatively, we could track a small amount, know who you are, and then actually get to know you. We are opting for the latter. It builds more trust and, from a business standpoint, it gives us better, more actionable feedback in a shorter space of time. It is truly a no brainer for us and for users.
Inevitably, there will be users who, for whatever reason, don't want their usage of Fig to be personally tracked. Of course. It's your data and software on your device. You shouldn't even need a reason. We want to cater to these users. And we will. But unfortunately, we are so early on in the process of building Fig that we need to be able to speak to our users. De-anonymising telemetry for the time being while minimisng the events tracked enables us to do this in a non intrusive way. But it is of course not for everyone. Therefore, once we reach a critical mass of usage in the coming months, we will then anonymise all telemetry, making Fig more accessible.
How does Fig load Autocomplete?
Our APIs read and write to the local shell. This means we get data (like what command you are typing), send it to our pre-loaded autocomplete app, and process it locally, then generate suggestions.
You can try this: with Fig running, try turning off your wifi. Autocomplete will still work.