Corvis LabsCorvis Labs
hubMIT Licensed · Free Forever

Open source is the foundation.

Everything we build starts in the open. The core tools are MIT-licensed and free forever. Enterprise is the production layer you add when you need it — MCP integrations, audit trails, tenant isolation, and real support.

terminal

agent-framework-cli — development workflow tool

Makes software development easier. Wraps GitHub Copilot in a structured 10-phase lifecycle with spec-first development, security gates, and 3 mandatory human approval steps. The tool that governs how your team builds with AI.

hub

foundry-kit — agent deployment framework

Builds production-ready agents for enterprise internal tooling and client-facing applications. Combines Azure AI Foundry, AI Search, document ETL, and ToolBuilder into one Python SDK that deploys to Azure Functions. The framework you build with, not just use.

terminal

agent-framework-cli

MIT

Governed AI for software development

Wraps GitHub Copilot in a structured 10-phase lifecycle. Spec before code. Architecture review before implementation. Security gates before merge. Five specialized agents that actually collaborate.

  • done10-phase governed lifecycle (Specify → Finish)
  • done5 specialized agents: orchestrator, architect, security, dev, QA
  • done3 mandatory human approval gates
  • doneGit-based or Beads persistent memory
  • doneWorks with any GitHub Copilot subscription
terminal
$ npm install -g agent-framework-cli
✓ installed successfully
$ acli setup && /acli.run "Add rate limiting to payments API"
Running...
upgrade

Enterprise layer adds MCP integration, Azure tenant deployment, audit trails, SSO, and production support.

hub

foundry-kit

MIT

Framework for building deployment-ready agents

A Python framework for building production-ready AI agents on Azure — for enterprise internal tooling and client-facing applications. Wire Azure AI Foundry, AI Search vector indexes, custom tools, and document pipelines into a single ergonomic SDK that ships to Azure Functions with one command.

  • doneDocument ETL helper — ingest PDF, DOCX, XLSX into Azure AI Search
  • doneAI search helper — vector + semantic RAG with Azure AI Search
  • doneFull Azure AI Foundry integration — agents, threads, runs
  • doneToolBuilder SDK — define function, API, and search tools with zero boilerplate
  • doneOne-command deploy to Azure Functions
terminal
$ pip install foundry-kit
✓ installed successfully
$ foundry-kit agent init my-bot --model gpt-4o
Running...
upgrade

Enterprise layer adds MCP integration, Azure tenant deployment, audit trails, SSO, and production support.

favorite

Why open source?

We believe the best developer tools get better when the community can see the source, file issues, and contribute. The core of what we build will always be open. Enterprise features are the business layer — not a paywall on the fundamentals.