LATEST CODE HERE: https://www.rhett.blog/projects/gemini-api/
Table of Contents*
- Fee Structure
- Set Up Gemini API
- Buying With AWS Lambda Functions
- Automation With AWS Cloudwatch
- Final Thoughts
*Nothing I post is financial advice
♊️ Sign Up for Gemini and get $10 of BTC when you Trade $100: https://gemini.com/share/ZVG91T
🎥 Full Video Tutorial (if you don’t like reading): https://www.youtube.com/watch?v=h6r1h3am6kA
💸 Fee Structure 💸
The reason that Gemini is the best exchange to use is that WHEN USED CORRECTLY they have the best fee structure
However, there are lots of way to mess this up. Example 1 (that most people will use) is the mobile fee structure.
You can see here when you do a regular mobile order (which is very convenient and likely how most people will by default use Gemini) the fees can be anywhere from 1.49% to 10% (!!) – this is WAYYYYY too high
Where Gemini excels is their API Fee schedule + their Transfer Fee schedule (don’t snooze on me here just because I brought up APIs – I’ve already coded everything for you)
You can see here at the lowest volume tier they offer .35% taker fees, but we can do better than that. We’ll be shooting for that .1% maker fee (100X better than the worst case 10% fee on the mobile schedule).
Some other exchanges (Kraken / Binance) also have .1% fees and in some cases (like Binance) allow you to get even lower when you pay the fees using their token.
The reason Gemini is better than both of these exchanges from a fee perspective is that Gemini doesn’t charge you on chain withdrawal fees on your first 10 withdrawals each month. This is huge for tokens like Ether and other ERC-20 tokens with massive gas fees, but it will save you money on Bitcoin too.
♊︎ Set Up Gemini API ♊︎
To take advantage of these low fees we need to set up access to your Gemini API. So after you’ve set up a Gemini Account here you’ll head into settings -> API (Programmatic access)
And you should see a screen that looks something like this
Click that big CREATE A NEW API KEY button
And copy your API key / API Secret into a safe place (and make sure never to show them to anyone or they will be able to trade/transfer on your behalf) the API Keys above are for demo purposes only and are not active
Select fund management and trading settings like we have in the screenshot above
Confirm that you’ve copied and pasted your API Key (Public Key) and API Secret (Private Key) into a safe place we’ll be accessing them in the next section (if you ever lose them or the script stops working for some reason you’ll need to come back here and regenerate new ones – no big deal)
Hit confirm and you’re done! – Pretty easy
🤑 Buying With AWS Lambda Functions 🤑
First you’ll need to make an AWS account (you might need to provide them credit card info but all the services we’re using are completely free as of this writing – you can check the current cost of Lambda Functions and Cloudwatch here)
If you’re like “But Rhett I don’t wanna make an AWS account”
It’s 2021! The world is teeming with FREE opportunity!
AWS is one of those opportunities.
Learn it, level up your brain, and prosper!
Now that you’ve set up AWS, all the code you’ll need is hosted for you here in Notion
The first thing you’ll want to do is download this gemini-layer.zip file from the Notion link. This zip file contains all the Python dependencies for the code hosted on the Notion page (namely gemini-python and gsheets [Google Sheets])
If you are suspicious of the code in the layer.zip feel free to generate your own. Instructions here to generate the layer zip using Docker + your own local Python environment.
Once you’re in AWS you’ll head over to the Management console and you’ll go over to Lambda under Compute
Inside Lambda head over to Layers on the left nav and hit Create Layer (top right)
Inside Create Layer, name your layer and upload the gemini-layer.zip file from Notion and choose Python 3.8 as the runtime
Hit create and then head over the functions in the left nav of Lambda and hit create function
Choose Author from Scratch, name your function, and choose Python 3.8 runtime
Delete the code in lambda_function.py and paste in the code from the Notion link above
Fill in the public_key and private_key strings with your public key (API Key) and private key (API Secret) from the Gemini API step above.
It should look something like this (yours will obviously be different but format should be the same):
public_key = “account-SDcdX7y3E5XsYRy6pkt8”
private_key = “AytmxmozZbeNokpfnRFcZfYkrq”
Next go down to add a layer (bottom right of screenshot)
Choose custom layer and pick the layer you imported earlier in this step from gemini-layer.zip
After you hit add, come back to the code tab hit the deploy button to save your changes.
Note that by default on line 21 of the buy-bitcoin the script will try to buy $20 of Bitcoin. You can change this number to whatever number you want.
When you hit test, you can fill out an event name and hit create (leave the code alone).
And then your script will try to buy $20 of Bitcoin from your Gemini account (note: you obviously have to have more than $20 in your account for this to work).
If you are getting timeout issues go to “Configuration” next to the code tab and hit edit on general configuration and change the timeout from 3 seconds to 30 seconds.
You should now have a script that allows you to buy $X from Gemini, so next let’s automate it so you can buy once a day/week/month/year etc.
🤖 Automation With AWS Cloudwatch 🤖
Go back to the AWS Management Console and click on CloudWatch
Inside CloudWatch navigate to the Rules tab under Events in the left nav
Click Create Rule here
Now click into Schedule instead of Event Pattern under Event Source
You can either change the fixed rate of days to the number of days you want or you can get fancy with Cron expressions
Cron expressions will allow you to set your own custom timing for how often these buy / transfer / sheet updating script runs.
For more info on Cron timings check out Amazon’s Documentation here
Some popular Cron timings:
15th + 30th of each Month: 0 12 15,30 * ? *
Every Friday: 0 12 ? * FRI *
Last day of each Month: 0 12 L * ? *
Once you’ve set your timing add your lambda function as the target on the right side under Targets
Configure details, give it a name, AND YOU’RE DONE!!!
🧠 Final Thoughts 🧠
There are some other little things you’ll need to do for the transfer / google sheets updating scripts
In the transfer script you’ll need to whitelist the address within Gemini in Settings and provide your BTC address(es) in the script – same way we included the public and private keys here earlier
The Google Sheets script is a little more complicated you need to set up your Sheets API – but it is really worthwhile because it will give you ultimate control in knowing how much money you’re actually making with this BTC dollar cost averaging strategy.
Video Walkthrough of the Google Sheets Script setup here
Again none of this is financial advice. This strategy has worked out really well for me, and the automation has been really powerful in allowing me to stop worrying about day to day price fluctuations so I wanted to share this with others.
If you do get this set up, let me know. If you have questions or get stuck feel free to leave them below or on the YouTube video above and I’ll do my best to get back to you A$AP Rocky
Happy Stacking 🙂