Our goal is to make collaborative software development scale without losing any efficiency. As teams get larger, it becomes challenging to maintain a consistent & efficient recipe for software development. Code reviews are one of the main checkpoints and safeguards for team alignment today. With code generation getting commoditized, we’ll need to allocate more efforts on the alignment layer, which is where we step in.

​
What we handle for you

Preparing your team for the next era of software engineering:

Engineering consistency

Creating a requestable representative for your engineering culture.

Team-specific coding assistance

Providing project-specific coding assistance.

Portable context for developer tools

Powering next generation developer tools with this same alignment.

​
What is left up to you

You make the decisions, we handle the compliance:
  • Decide the modifications you’d like to see in the way you produce software
  • Allocate resources to produce code (written by humans, or generated by a model)
  • Orchestrating the different parts of your software development workflow, and assign them the most relevant context

​
How it works

The current version of the system relies on two main pipelines that use deep learning models (namely language models, either instruction-tuned or completion-based).
From an architecture perspective now, here are all the system components:

​
Server resources

Guideline database

A PostgreSQL database to store and edit the coding guidelines of each project, hosted by our friends from Supabase.

Deep learning models

Either our cloud-hosted or locally-run models with coding capabilities, served by Ollama.

Backend API

The gatekeeper to the database and computing resources that manages access. Built with FastAPI

API monitoring

Services to run analytics (PostHog), error monitoring (Sentry), and notifications (Slack).

​
Client applications

VSCode extension

Smart linting & code chat with relevant context for developers.

Quack chat

ChatGPT but for coding.