Panacea AI Coding Toolchain¶
This recipe explains how Panacea’s AI coding experience is delivered across CLI, SDK, and VS Code.
What you'll learn¶
- The different AI coding entry points in Panacea
- How the CLI, SDK, and VS Code extension relate to the shared backend
- Key product capabilities for private code assistance
- Where to look in the repo for implementation details
Why this matters¶
Panacea is built as a unified product with multiple interfaces:
- a CLI that powers
anote chat, code search, and repo review - a VS Code extension for in-editor AI assistance
- a SDK for embedding Panacea into other applications
These interfaces share a backend and agent-driven reasoning layer, which makes the product consistent across desktop, web, and code workflows.
Key Panacea files¶
| File | Why it matters |
|---|---|
Panacea/packages/cli |
TypeScript CLI implementation for developer workflows |
Panacea/packages/vscode |
VS Code extension and chat integration |
Panacea/packages/sdk |
TypeScript SDK for programmatic access |
Panacea/packages/backend |
Shared backend service powering all UI and CLI interactions |
How it works¶
- A coding user action starts at the CLI, SDK, or VS Code extension.
- The request is sent to Panacea’s backend API.
- The backend uses agent orchestration and model providers to produce code-aware answers.
- The response is returned in the same interface, with code suggestions, explanations, or fixes.
Useful product features¶
- CLI:
anote chat, repo search, code review, code generation, and embeddings-powered assistance. - VS Code: inline chat, diff previews, code actions, and streaming responses.
- SDK: a client wrapper for the Panacea API, enabling custom integrations.
Run it locally¶
From the workspace root (anote/panacea):
bash
cd Panacea
cp packages/backend/.env.example packages/backend/.env
docker compose up --build
This starts the shared backend and frontend services.
In another terminal, run the CLI package:
bash
cd Panacea/packages/cli
npm install
npm run dev
Then you can use the CLI locally or build it with npm run build.
For VS Code development, open Panacea/packages/vscode in VS Code and launch the extension with the debugger.
Notes for the cookbook¶
This recipe is useful for teammates who need a high-level walkthrough of Panacea’s multi-interface AI coding product. It can also point readers to implementation files they can modify or extend.