Getting started with OpenProtein.AI’s API#

Step 1: Request early access

Step 2: Install our Python client

You can install the package via pip or conda as below:

pip

pip install openprotein-python

conda

conda install -c openprotein openprotein-python

Github

The source code is available here.

Want to start with the web version?

Visit Get started with no code

Step 3: Authenticate your session

Use your username and password credentials generated at sign-up to authenticate your connection to OpenProtein.AI’s backend.

import openprotein

with open('secrets.config', 'r') as f:
   config = json.load(f)

session = openprotein.connect(username= config['username'], password= config['password'])

OpenProtein Job System

The OpenProtein.AI platform uses a job system to support asynchronous task execution. Upon initiating a task, the system will schedule a job with a unique Job ID so you can return at a later time for tasks with long processing times.

OpenProtein API session

Executing workflows is achieved with the OpenProtein APISession object (see openprotein.APISession())

session = openprotein.connect(username="username", password="password")

You then have access to all the workflows: For example

session.data.create()

Or

session.poet.create_msa()

Step 4: Get started using our API for your protein engineering goals

Quick start tips#

Do you want to…

Make sequence predictions or designs without using any data?
Get started with PoET which uses evolutionary information to generate protein sequences
Analyze your experimental data for library design?
Use our Property Regression Models to train and deploy machine learning models in your context
Explore your protein's structure?
Use our Structure Prediction workflow to obtain PDB files using ESMFold and AlphaFold2 models
Obtain embeddings from protein language models?
Use our Foundation Models to access high quality sequence embeddings using proprietary and open source models
On this page