Ariana is a CLI to automatically add observability to your code and an IDE extension to consume it & provide context-aware debugging capabilities to coding agents. You don't have to change any code in your codebase or specify breakpoints. Currently supports JS/TS & Python. ✨ Key FeaturesUse Ariana VSCode extension to :
💾 How to install
🧵 How to useFollow the Getting started instructions in the Ariana extension panel for the most up-to-date guidance. Below is a summary: 1. Install the
|
Package Manager | Command |
---|---|
npm | npm install -g ariana |
pip | pip install ariana |
python -m pip install ariana |
|
python3 -m pip install ariana |
2. Observe your code with ariana
To collect runtime information, run your usual build/execution command, but prefix it with ariana
. This tells the CLI to instrument a copy of your code (in a local .ariana
directory) and observe its execution.
ariana <your usual build & run command>
For example:
Codebase Type | Command |
---|---|
JS/TS | ariana npm run dev |
Python | ariana python myscript.py --some-options |
Run this in each terminal where you execute a part of your application you want to observe.
3. View Traces in VS Code
Open the Ariana panel by clicking on its icon in the Activity Bar.
- If you've run
ariana
on multiple projects, you might be prompted to select the run you want to focus on. - Traces from your code's execution will appear in the Traces tab.
4. Analyze Traces & Hover Over Code
Once traces are loaded, Ariana provides insights directly in your editor:
🗺️ Execution Highlighting: See which parts of your code ran. | Highlight Color | Meaning | |-----------------|-------------------------------------------------| | 🟢 Green | Code segment ran successfully. | | 🔴 Red | Code crashed here. | | ⚫ Grey/None | Code segment didn’t run or couldn't be recorded. |
🕵️ Value Hovers: Hover over any expression in your code to see its last recorded values and execution time.
5. (Optional) Use AI to Understand Traces (WIP)
In the Ariana panel, you can copy the collected traces. You can then paste these into an AI coding assistant with a prompt like:
<paste traces>
Using the debugging traces above and your knowledge of the codebase please do <xyz>
Tip: Use an agent with a large context window, as traces can be verbose.
Coming soon: More compact traces and an integrated AI agent for runtime analysis and fixes.
Preview :
To test Ariana before using it on your own code:
git clone https://github.com/dedale-dev/node-hello.git
cd node-hello
npm i
ariana npm run start
Troubleshooting / Help
😵💫 Ran into an issue? Need help? Shoot us an issue on GitHub or join our Discord community to get help!
Requirements
For JavaScript/TypeScript
- A JS/TS node.js/browser codebase with a
package.json
- The
ariana
command installed withnpm install -g ariana
(or any other installation method)
For Python
- Some Python
>= 3.9
code files (Notebooks not supported yet) - The
ariana
command installed withpip install ariana
outside of a virtual environment (or any other installation method)
Supported languages/tech
Language | Platform/Framework | Status |
---|---|---|
JavaScript/TypeScript | Node.js | ✅ Supported |
Bun | ✅ Supported | |
Deno | ⚗️ Might work | |
Browser Frameworks | ||
JavaScript/TypeScript | React & .jsx / .tsx |
✅ Supported |
JQuery/Vanilla JS | ✅ Supported | |
Vue/Svelte/Angular | ❌ Only .js / .ts |
|
Other Languages | ||
Python | Scripts / Codebases | ✅ Supported |
Jupyter Notebooks | ❌ Not supported (yet) |
Code processing disclaimer
We process and temporarily store for 48 hours your code files on our server based in EU in order to instrument them and help you debug afterwards. It is not sent to any third-party including any LLM provider. An enterprise plan will come later with enterprise-grade security and compliance. If that is important to you, please let us know.
Licence
Code generated and/or transformed by Ariana is yours and not concerned by the following licence and terms.
Ariana is released under AGPLv3. See LICENCE.txt for more details.