Open Source - supercog-ai/community GitHub Wiki

The Supercog code is not yet open source. We are big believers in the cloud - and we wanted to provide the easiest path to getting started with Supercog. As such we decided to focus on getting our hosted version running first.

That said, there are lots of advantages to running the software yourself, and we want people to have that opportunity. So our roadmap looks like this:

1. Run via Docker. We already build containers. We need to tighten up some of the configuration, write some docs, and then just publish our containers to a public registry. We expect this to be a low effort way to run locally.

2. Open source our agent tools. This is the primary place that folks are likely to want to extend Supercog.

3. Enable running your agents on prem. We think running agents locally (on your laptop or in your VPC) will be a popular option for privacy and security reasons. Our plan is to let you use our hosted Dashboard, but connect your own instance of the agent runner.

4. Bring your own LLM. We have a pluggable interface for LLMs already, and we plan to allow people to bring their own pre-trained or licensed LLM to run their agents.