LiveFlow runs your actual Python, JavaScript, or TypeScript project, multiple files, cross-file imports, and interactive input() included, right in your browser. Press run and watch every value, call, and collection light up as nodes you can follow, then scrub back to the exact moment anything went wrong. No install. No print statements.
Type Python, JavaScript, or TypeScript, single file or a whole multi-file project, and hit run. No install, no setup. It runs right in your browser.
Every variable, function call, and list becomes a node. Data flows in on cyan, the path of execution flows out on amber.
Drag the timeline and the whole picture rebuilds at that exact step, so you can see what changed, and why.
1from questions import QUESTIONS, check_answer23def run_quiz():4 score = 05 for question, answer in QUESTIONS:6 reply = input(question)7 if check_answer(reply, answer):8 score += 19 print(f"You scored {score}")
Whether you’re learning an algorithm, finding your way around an unfamiliar codebase, or checking code an AI just wrote, it comes down to one thing: actually seeing what runs.
Sorting, recursion, linked lists, trees: stop tracing them on paper. Run your algorithm and watch the array fill, the recursion unwind, the pointers move. The “aha” lands faster when you can see it.
Dropped into an unfamiliar codebase? Run it and watch the call flow light up: which function calls which, and what data moves between them. When you’re ready, pull the same traces into your own tools through our API.
Got code from an AI assistant? Run it here first, somewhere safe, and see exactly what it does before it touches your project. Every run is saved as a clean recording that you, and your AI, can read back.
When an AI writes code, the hard part is trusting it. LiveFlow runs that code safely and records everything that happened, so both you and your coding agent can see what really ran, spot the bug, and fix it.
AI-written code runs on its own, with time and memory limits, so it can’t touch your machine or your files.
Every run becomes a clean, line-by-line trace an agent can parse to find exactly where things went wrong.
Save any run as a snippet: a link that replays the exact execution, no rerun required.
Send code, get back a trace. Wire LiveFlow straight into your own agent loop or developer tool.
1{ "event": "FUNCTION_CALL", "fn": "run_quiz", "file": "main.py" }2{ "event": "CONSOLE_OUTPUT", "text": "What planet do we live on? " }3{ "event": "FUNCTION_CALL", "fn": "check_answer", "file": "questions.py" }4{ "event": "STATE_DELTA", "name": "score", "value": 1 }
A real run, replayable and machine-readable.
Watch control move through your program. Every call, branch, and loop draws itself as it happens, in amber.
See every variable, in every scope, at every step. Lists and objects show what changed: what was added, sorted, or removed.
Jump to any step and the whole picture rebuilds exactly as it was, the same way, every time.
Every plan runs zero-setup in your browser, with live input() and full timeline scrubbing. The Free plan needs no card: 15 runs a day in one language you choose, resets daily. Paid plans unlock all three languages, longer runs, multi-file projects with cross-file imports traced automatically, and the libraries you actually reach for, like NumPy, pandas, lodash and more, preinstalled and ready to import. Each starts with a 7-day free trial of that plan (cancel anytime). Billed in INR via Razorpay; students at partner colleges get a free first month, then ₹10/mo. See the full library list.
Student plan is for verified students at allowlisted colleges. Eligibility is automatic once you sign in with your institutional email (no extra steps). Not listed? Ask us to add your college.