​
Integrations

Users generally start with GitHub and Teams integrations.

These two sources provide a great base for Asimov, but we highly encourage providing additional sources of context. Reflection is continuously adding integrations as sources to expand the reach and power of Asimov.

One of the most powerful features of Asimov is its ability to combine its long context capabilities with external knowledge sources. The more codebase context Asimov has, the more accurate its responses and actions.

We’re actively expanding our integration library. If there’s a tool that’s critical to Asimov deeply understanding your codebase, let us know. We prioritize new integrations based on user feedback and can fast-track high-impact connections, including support for internal or home-grown systems. Email support@reflection.ai to request an integration.

​
Prompts our users love

Below are the most popular initial prompts our users and team love:

  • Review our Teams interactions and derive an FAQ that can usefully onboard new engineers to frequently-asked questions in and around our code.
  • In service of rapidly onboarding new engineers, create a tree-like graph that describes our codebase, with a short description annotating each node.
  • Analyze our codebase for performance bottlenecks or inefficient patterns. Highlight three specific areas where optimization could yield significant improvements in speed, resource usage, or responsiveness. Provide data or reasoning to support your suggestions.
  • What are the most common style guide violations in the codebase?
  • Cross-reference our code comments with our documentation to identify areas where docs are outdated or missing critical information.
  • Analyze TODO/FIXME comments in code and correlate them with discussions in chat to prioritize which ones to address first.