Questions and answers

A new automation puzzle every week, built from a real question somebody asked in the n8n community. You get your own version of the data, you build an n8n workflow that solves it, and you submit the answer.

Yes. Registration stays open between challenges. Register once with your community username and keep the participant key sent to your forum inbox. Registration does not start your timer or enter you in a challenge. Data access, answer checks and submissions are available only during the published challenge window.

The challenge stays open through Saturday night and closes at 00:00 UTC on Sunday, the instant Saturday becomes Sunday in UTC. Challenge 02 closes at 2026-09-27 00:00 UTC. This is not Friday night or Sunday night. The exact UTC timestamp on the challenge page is authoritative. Any deadline change will be announced in the questions thread. Registration stays open after participation closes.

No - and for Photo Finish you must not. Submitted workflows may use only the published native node allowlist; Code, Function, Function Item, SQL Merge, code Sort, credentials, outside HTTP URLs and expression helpers such as map/reduce are refused before grading. The exact reviewed versions and expression rules are on the challenge page. AI help is allowed, but points are held as pending replay until the workflow computes both hidden replay worlds after the close.

Yes. The challenges are built from questions real users asked, which means they are the kind of problem you hit in normal work rather than a puzzle for its own sake.

Create an account, get your participant key by forum private message, read the challenge, build a workflow, submit your answer. Create your account with your community username and an email.

So the data can be yours. Everyone gets a different dataset generated from their username, which is what makes the answers un-copyable and lets us tell you apart on the leaderboard.

No.

A short code shaped like MB-XXXX-XXXX. It is how the site knows a request is really from you. Personal challenge requests need it - fetching your data, opening a stage file, checking an answer and submitting. Your username alone will not do, because usernames are public: they are on the leaderboard. It is the only credential in the whole programme - there is no password. Public pages and the free practice routes do not need a key.

It is sent to you as a private message on community.n8n.io the moment you register. We never show it on a page and never put it in a link we send you - not at registration, not when you ask for a new one. The forum message is the only place it exists, which is also how we know the account is really yours. You will put it in your own requests - see "How do I send my key?" below.

Ask for a new one at the reset page. It sends a fresh key to the same forum account. The old key is replaced only after the forum accepts the private message and the reset is saved. If delivery fails, your old key still works.

Treat it like one. Anyone holding it could submit as you, and - because fetching your data is what starts your clock - could start your clock while you are asleep and wreck your ranking. It cannot touch your forum account and it does nothing outside this programme, but do not paste it into a public thread, a screenshot or a shared workflow. If you do, reset it immediately.

Use a header: x-participant-key: MB-XXXX-XXXX, which is what the starter workflow uses. Do not put the key in a URL: query strings can land in browser history, screenshots, proxy logs and monitoring systems. The web forms collect it in their private POST body. Treat the key like a password.

A missing key gives you 401 with a reminder of how to send it. A key that does not match that username gives you 403. Neither one starts your clock, costs you an attempt, or counts against a rate limit, so a misconfigured workflow cannot hurt your ranking - fix it and carry on.

Your workflow JSON is not published. It is stored only for a correct submission and is available to organisers for review. The submitted participant key is redacted before storage, and a "key redacted" note confirms when this happens. Credential references are also removed. This does not detect every hardcoded secret: remove all keys and tokens yourself before submitting. If a key has been exposed, reset it.

Please do not. One entry per person, and a submission made with your key counts as yours.

Check your forum inbox rather than your email, since it is a forum PM. Almost every time, the username typed at registration is not an existing community.n8n.io account or is spelled differently there - and the registration page says so when the message could not be sent. A live participant key is never shown on a page or placed in a generated link, so there is nothing to copy from the screen. Check the spelling on your forum profile, then register again with the same username and the same email: you get the same key back and we send the message again.

On purpose. Your dataset is generated from your username, so no two participants get the same answer. It means you cannot copy a result, and it means somebody can post their whole workflow after the deadline without spoiling it for anyone.

Not usefully. The data endpoints are rate limited, and the wrong answers are all plausible - nothing in the response tells you whether you are on the right track. Working it out is faster than guessing.

Requests that change an account are processed one at a time to keep counters, start times and submissions consistent. If two overlap, the second gets 409 Conflict with a short retry delay. Wait a moment and retry it. That refused request spends no budget and does not replace your submission. Different accounts can still use the challenge at the same time. Registration has its own short shared reservation so overlapping sign-ups cannot create conflicting accounts; a busy registration can also be retried. If it keeps happening after your other requests have finished, contact the organisers.

Photo Finish allows 120 data fetches per rolling hour, shared across the data files. Each stage allows 4 checks per 15 minutes and 10 well-formed checks for the whole challenge; malformed values are free. Submit accepts at most 30 requests per hour, and only 10 graded wrong submissions are reserved for the challenge. Correct entries are recorded as pending replay and earn no points until the isolated replay passes. The public practice route contains four small, non-chainable teaching lessons; the full race mirror is private QA only.

Yes. Claude, ChatGPT, the n8n AI builder - all fine. You do have to explain your workflow in your own words, so make sure you understand what you built.

Yes. Mention it in your explanation.

Talk about the approach as much as you like. Submit individually - you each have different data anyway, so you would each need to run it yourself.

Please do not post a full solution publicly until results are out. Afterwards, share it everywhere - that is the point of the programme.

At the submit page, linked from every page on the site. That is the only step - you submit your answer there directly.

Always: your community username, your participant key, your answer, a short explanation in your own words, and your workflow JSON. Plus whatever this week's challenge needs - the fields change with the challenge, and the page always shows the current ones.

Yes, while the challenge is open. A later wrong submission will not replace an earlier correct solve. There is a limit of 30 submissions an hour so that the page cannot be used to guess an answer, which is far more than resubmitting honestly needs.

No. Once you have a correct answer on record it is never downgraded by a later wrong submission. This does not exempt an entry from workflow review or disqualification. Your original solve time is kept, so your position on the leaderboard does not move, and a later wrong answer cannot undo a correct one. You can resubmit purely to improve your explanation.

Yes. The answer alone does not show that you built anything, and a workflow is the whole point of the exercise. It is only stored when your answer is correct - a wrong submission never leaves a workflow behind. There is no link field any more: a link rots, and the JSON is the actual artefact.

You can still take part, but not with an empty submission - the JSON field is required. If you worked the answer out by hand or with an AI, build the workflow that produces it and send that. Entries whose workflow was clearly never run are not eligible for a prize, and they are checked by a human before anything is awarded.

Credential blocks are stripped from every node before it is stored, so credential names and ids do not survive. That cannot catch a token hardcoded into a URL or a header, so remove secrets yourself before pasting. There is a 60KB limit.

So nobody can copy, and so solving early is not punished. When the challenge closes, every correct entry - the explanation you wrote, in your own words - becomes visible on that person's public record at the same moment. Your workflow JSON is not part of that: it stays private to you and the organisers either way.

The status page shows exactly what is on record for you, including whether it was marked correct.

Automatically, the moment you submit. The whole challenge is re-derived on the server from your username, so correctness is never taken from anything you send.

By how long you took, not by when you finished. Your clock starts the first time you fetch your own challenge data and stops at your first correct submission. Checking a stage answer alone does not stop the clock. Only correct entries are ranked. See it at the leaderboard.

The first time you fetch your own data from the challenge endpoint. It is written once and never reset - fetching your data again does not restart it. So fetch it when you are ready to work. Reading the brief, thinking about it, or building a workflow beforehand all cost you nothing.

Because the challenge runs across every timezone. Under the old rule, somebody whose morning lines up with the Monday release could not be beaten however good anyone else was, and anybody arriving on Wednesday found the board already settled. Timing how long you took gives everyone the same clock whenever they start.

Your public profile lives at https://marketing.app.n8n.cloud/webhook/cwc/user?u=YOUR_USERNAME, and it shows your entries and any past challenges you have solved.

Challenge 01 awards 11 points for a correct solve, regardless of rank. There is no rank bonus for Challenge 01. Challenge 00 keeps its original rules: 10 points for a solve, +5 for 1st, +2 for 2nd or 3rd, +1 for anywhere else. For Challenge 00, a win is 15 and a finish is 11. A wrong answer or an unfinished attempt scores nothing. Your total carries from one challenge to the next and only resets when a new season starts, so the number beside your name on the leaderboard is everything you have earned this season.

Challenge 01 rewards finishing the work: a correct solve earns 11 points whether you finish first or later in the week. Rank still reflects elapsed time, but does not increase this challenge's points. Challenge 00 retains its original rank bonuses.

Yes. Your total is kept when a challenge closes and carries into the next one - it does not go back to zero every Monday. A season is four challenges, and only a new season starts everybody from nothing again. On the leaderboard that is a single Points number beside your name - everything you have banked this season, plus what this week is currently worth. One number, not a running total and a weekly one.

Closing and archiving preserve the result and its challenge-specific scoring rule. Challenge 01 awards a flat 11 points, so changes in rank do not change its solve points. Results remain subject to workflow review, disqualification and logged corrections, including after closing. New challenge rules do not rescore Challenge 00.

The one with the lower total elapsed time across the season - the same clock that ranks each week. Somebody whose solve was never timed sorts last on that tiebreak rather than first.

They can be corrected, and every correction is logged with a reason. It is there for the cases a formula cannot settle - a result that turned out to be wrong, a week where an endpoint of ours broke, a placing decided by reading two workflows rather than by the clock. Your rank, your time and your workflow are never rewritten by it; only what the result is worth changes, and your record still shows where you actually finished.

Yes. A correct answer is not automatic approval of the submitted workflow. Organisers may inspect and rerun a submission to check that it works and produces the claimed answers for that participant. Results and prize eligibility remain subject to this review, including after the challenge closes.

An entry may be rejected or disqualified if its workflow cannot reproduce the submitted answers, uses a forbidden node, or there is verified evidence of cheating or abuse. Organisers may remove the affected result from eligibility or reduce its awarded points. Corrections are recorded with a reason. If you believe a decision is wrong, contact the organisers through the challenge thread and include the details needed to reproduce your result.

No. AI assistance is allowed. Review is based on the published rules and reproducible workflow behaviour, not a guess about who wrote it. For Challenge 01, the workflow must still contain no Code, Function or Function Item nodes and no Merge node in SQL mode. Automated structural checks do not prove that a workflow runs correctly, and an expression-heavy workflow is not by itself proof of cheating.

It costs you time on the board, yes. Your clock runs from your first fetch to your submission, and every hour in between counts - including the ones you spent asleep. There is no pause and no cap. That is the plain rule and it is the same for everybody. What it does not do is stop you finishing: solving it correctly on Thursday still counts as solving it, and the week stays open until it closes. If speed is what you are after, fetch your data when you are actually ready to sit down with it - reading the brief and planning your workflow beforehand costs you nothing, because none of that is timed.

The leaderboard and public profiles show participation and results while the challenge is open. Explanations for correct submissions become public after closing. Stored workflow JSON and submitted answers are not published. Results remain subject to organiser review.

Specific rewards and eligibility details will be announced separately. Participation does not guarantee a prize. The platform records challenge results and season points; it does not automatically award or draw prizes.

Anyone with a community.n8n.io account can take part, appear on the leaderboard and score points. Anything to do with rewards is decided by the people running the programme, and will be spelled out here before it applies to anybody.

No. Challenge 01 awards the same 11 solve points regardless of rank. You can practise building a workflow, keep a public participation record and discuss the challenge with the community. Rewards, if announced, have their own eligibility rules.

Always. Entering is never restricted, and your points keep accumulating.

Your community username, email, participant key and registration details; challenge start times, submissions and history; and operational records used for rate limits and organiser review. Public profiles do not expose your email or participant key.

Matching your account when you register again or request a key reset, and contacting you about a reward if one is arranged. Your participant key arrives by forum PM, not email. Your email is never published on a public profile.

You can check your submitted answer on the status page, and organisers can review it. Your username, participation, solve status, rank and points can appear publicly while the challenge is open. Your explanation stays hidden until closing; stored workflow JSON and submitted answers are not published.

It is stored only if your answer was correct, with credential blocks stripped. It is not published. The people running the programme read it to check entries and to settle a close finish, and it is deleted 28 days after the challenge closes. If we ever start publishing solved workflows so the community can read how people did it, that will be announced here first, and it will apply to challenges from then on - never retroactively.

Yes. Ask, and your participant row and submissions can be deleted. Results already published in a forum thread stay there, because they are part of the public record of that week.

Say so in the challenge thread. If it is a real outage the deadline is extended and that is announced publicly, not quietly.

Most of them are rolling windows, so waiting clears them: 60 an hour on stage fetches, 30 an hour on submissions, and 4 checks per 15 minutes on each stage of verify. One is not a window. The 15 properly formed checks per stage are a budget for the whole challenge and they do not refill, so waiting will not bring them back - but you can still work on the other stages, and you can still submit, because the submit page grades every answer together. If you burnt any checks on something that turned out to be broken on our side, say so in the thread: the rolling windows can be cleared for you and the spent checks can be handed back.

Tell us in the thread. If a challenge is broken it gets fixed, the deadline moves, and whoever spotted it is credited.

Post in the thread with what you got and how you got there - not the full workflow if the challenge is still open. Grading re-derives the whole challenge server side, so a genuine mistake is a bug worth finding.

Ask in the challenge thread on community.n8n.io first. Questions asked in the open help the next person who has them, and you usually get an answer faster.

Yes, for things that do not belong in a public thread - an account problem, a dispute about a result, a privacy request, or anything you would rather not post. Message Anshul Namdev on the forum. Please use the thread for ordinary questions - a direct message only reaches one person, and the answer helps nobody else.