Built for technical, live coding, and system design rounds. Capture LeetCode and CoderPad screenshots, drop in design diagrams, and watch an approach, complexity analysis, and working code stream token by token. Whisper Mode keeps the panel nearby while you work, with visibility depending on the meeting share source you choose.
Premium plan · 15 credits · 2-hour sessions
Approach
Two-pointer scan with a hash map of complements. O(n) time, O(n) space.
Solution
def two_sum(nums, target):
seen = {}
for i, n in enumerate(nums):
if target - n in seen:
return [seen[target - n], i]
seen[n] = iScreenshot · leetcode.png
Two Sum — Easy
Coding Interview is a Premium real-time copilot for the rounds where the question lives on a screen, not in a conversation. Capture a LeetCode or CoderPad screenshot, drag in a system design diagram, or attach a code file — and the AI streams an approach, complexity analysis, and working code (or a structured design review) into a panel beside your editor. Built for Brazil candidates interviewing at Big Tech, high-growth startups, and any company where the technical screen is the gate.
Coding Interview covers every technical round style — from verbal trade-offs to live coding on shared editors to whiteboard system design.
Concepts · debugging · trade-offs
Verbal Q&A on technical concepts, debugging scenarios, and applied design trade-offs. Mic + Auto-Detect listen to the call, transcribe questions, and stream structured answers — clarify, plan, implement, test — into the panel on the left. No code editor; you explain in your own words while Coding Interview suggests the structure.
LeetCode · CoderPad · HackerRank · CodeSignal
Screenshot the coding problem from any tool, stack up to 10 attachments, and add typed context (constraints, what the interviewer just said). The AI streams an approach, complexity notes, and a working solution in your chosen language into the AI panel on the left — adapt it in your editor as you talk.
Architecture · scaling · trade-offs
Drop a diagram, whiteboard photo, or architecture sketch into the Interview Feed alongside the prompt. The AI returns a structured review — components, data flow, bottlenecks, scaling, trade-offs — that you can use as a talking-point map while you walk the interviewer through your design.
Two panels — Interview Feed on the right, AI panel on the left. The session reads as a single conversation, even when you mix screenshots, code files, and dictation.
Set the language at the top of the session — the AI matches your syntax for every solution it streams back.
Click the camera icon to screenshot the problem, drop in a code file, or paste from the clipboard. Stack multiple captures into a single send (up to 10, 5 MB each).
Add typed context — constraints, the interviewer's exact wording, follow-ups — then send. Your prompt and attachments stay in the Interview Feed on the right.
The AI streams approach, complexity notes, and working code (or a design review) into the panel on the left. Read it, internalize it, and write the answer in the editor the interviewer is watching.
Read the streamed solution, then write the answer in your own words — and your own code — in the editor the interviewer is watching. Coding Interview is a prompt, not a script.
Vision, streaming, and a Document Picture-in-Picture overlay — combined into a single session view that mirrors how candidates work today.
Interview Feed on the right captures every prompt — typed, dictated, screenshotted, or uploaded. The AI panel on the left streams the analysis. Everything lives in one chronological transcript.
Send a LeetCode problem image, a CoderPad screenshot, or a whiteboard photo. The vision model parses code, constraints, and architecture sketches and reasons over them directly.
Solutions and design reviews stream over SSE so you see the AI thinking in real time — no spinner, no full-page reload, no waiting until the very end.
A Document Picture-in-Picture overlay keeps the AI panel visible while you work. It is designed for private notes during single-tab or single-window meeting shares. Entire-screen sharing can expose anything on your display, so test your setup first.
Up to 10 attachments per send — screenshots, code files (.js, .ts, .py, .java, …), or design images. Stack them when the problem spans multiple screens or files.
On End Session everything is merged — typed prompts, dictated questions, screenshot captions, code files, diagrams — into one review with AI scoring across Communication, Technical, Problem Solving, and Confidence.
Live Coding and System Design both support Whisper Mode — a 340×480 always-on-top overlay built on Chrome's Document Picture-in-Picture API. The same React panel is portaled into that window so it stays in sync as the AI streams.
Coding Interview's audio capture and your meeting's screen share are separate choices. In the meeting, share a single tab or window when appropriate instead of the full display. Entire-screen sharing can expose anything visible on your screen, so run a test call and confirm the interview's policy on outside assistance before the session.
Interview Copilot listens to live call audio for behavioral, case-study, and phone interviews — the questions are spoken. Coding Interview is for rounds where the question lives in a code editor or whiteboard — you capture it via screenshot or upload, and the AI returns code or a design review instead of conversational guidance.
Coding Interview uses heavier real-time analysis — vision for screenshots and diagrams, code understanding, and design reasoning. It is included on the Premium plan; Interview Copilot remains available on Pro.
In a Live Coding session, click the camera icon in the Interview Feed to screenshot the problem from LeetCode, CoderPad, HackerRank, or CodeSignal, or click the paperclip to attach a code file. Stack multiple screenshots in one send if the problem spans screens. Type any extra context in the box and press send — the AI streams an approach, complexity notes, and working code into the panel on the left.
Yes. In a System Design session, drop a diagram, whiteboard photo, or architecture sketch into the Interview Feed (drag-and-drop or paperclip). Add a short prompt and send — the AI streams a component-level review covering data flow, bottlenecks, scaling, and trade-offs.
Technical Interview uses live audio capture — mic + Auto-Detect listen to the call, transcribe questions, and stream answers into the AI panel. There is no code editor; the view is optimized for verbal Q&A on concepts, debugging, trade-offs, and applied scenarios. Same browser caveats as Interview Copilot (Chrome or Edge desktop).
Whisper Mode is designed for private notes during single-tab or single-window meeting shares in Zoom, Meet, and Teams. Entire-screen sharing can expose anything on your display, so use a narrow share source and test your setup before the interview. The two operations are independent: what you share with Coding Interview for audio capture is different from what you share with the interviewer.
A Coding Interview session is 15 credits and lasts up to 2 hours — the same envelope as Interview Copilot.
Live audio capture (Whisper Mode + Auto-Detect) requires Chrome or Edge on desktop. Live Coding and System Design upload flows work in Safari and Firefox too, with screenshot capture limited on those browsers and on mobile. Manual mode (typing or pasting) works everywhere.
Screenshot the problem, drag in a diagram, dictate a follow-up — and watch a working solution stream in beside your editor.
Live audio assistance for behavioral, case-study, and phone rounds — private in your workspace via Whisper Mode.
Learn morePractice technical and behavioral rounds with AI feedback before the real call so the live session feels familiar.
Learn moreStructured resumes reviewed against each target role, paired with Coding Interview practice and live assistance.
Learn more