Everything about Jellar
Your complete guide to building, testing, and earning on Jellar.
What is Jellar?
Jellar is a Human Signal testing platform for AI builders - people who use AI to build apps, tools, and products fast (sometimes called "vibe coders").
The problem: you can build something in a weekend with AI, but getting real humans to actually test it is hard. Friends are too polite. Social media is too noisy. User research firms are too expensive.
Jellar solves this. Builders list their projects, Testers review them with structured feedback, and everyone earns karma and credits for participating.
How it works
A builder lists their project with a URL, a short mission for testers ("try to sign up and send a message"), and answers a safety questionnaire. Automated checks run instantly, then a human admin reviews within 24 hours.
Approved projects appear on the Pulse globe and the Ladder. They're ranked by tier, then by builder karma within each tier. Sponsored projects sit at the top in gold.
Testers pick a project, read the briefing, then click to open it in a new tab. A 3-minute timer enforces minimum dwell time. Testers fill in structured feedback - friction points, magic moments, a summary.
Testers earn karma based on completeness. Builders see reviews in their dashboard. Builder karma improves, lifting their project higher on the Ladder.
Builders vs Testers
Builders
You've made something and want real feedback. You submit your project, define a mission for testers, and earn builder karma as reviews come in. Higher builder karma = higher position on the Ladder = more visibility = more reviews.
Testers
You spend 5–60 minutes using someone's build and give structured, honest feedback. You earn karma for every completed review. High tester karma gets you on the leaderboard.
Most people are both. Builders are encouraged (and on some tiers, required) to test other projects before listing their own. This keeps the community healthy.
Submitting a project
- Sign in with GitHub. Your account must be at least 5 days old.
- Go to Submit and fill in the form. You'll need a live HTTPS URL, a category, a mission brief, and answers to the Vibe Check safety questions.
- Automated pre-flight checks run instantly in your browser. If any fail, you'll see exactly why and how to fix it.
- If all checks pass, your submission enters the manual review queue. A human reviews it within 24 hours.
- Approved → your project goes live on the Pulse globe and Ladder. Rejected → you'll see the reason and can resubmit after fixing the issue.
Missions are currently capped at 60 minutes. If your project needs a longer test (e.g. tracking something over days), that requires a paid bounty mission - coming soon. For now, design your mission to fit within an hour.
Project categories
Pick the category that best describes what your project does, not what technology it uses. Not sure? If AI is the whole point of the product, pick AI Tool. If AI is just under the hood, pick the category that describes the user-facing purpose.
AI or an LLM is the core feature - chat assistants, AI writers, code generators, image tools, AI search.
Built for developers. APIs, CLIs, code helpers, debugging tools, deployment dashboards, devex improvements.
Helps people get things done. Task management, note-taking, workflow automation, scheduling, focus tools.
B2B or business-focused software. CRMs, dashboards, reporting tools, team collaboration, ops software.
Tools for making things look or sound good. Image generation, video editing, typography, brand tools, music.
Money-related tools. Budgeting, invoicing, expense tracking, payments, crypto, personal finance.
Fitness, mental health, habit tracking, nutrition, sleep, meditation, and wellbeing tools.
Tools that teach or help people learn. Courses, tutoring, flashcards, skill trackers, language learning.
Selling things online. Shops, marketplaces, product discovery, inventory tools, dropshipping.
Games of any kind, interactive experiences, browser toys, party games, entertainment apps.
Tools for talking to people. Chat apps, email clients, team tools, newsletters, social features.
Tools for understanding data. Dashboards, reporting, data viz, analytics, business intelligence.
Tiers explained
Your tier determines how many reviews your project can receive, how long your listing runs, and where you appear on the Ladder. Within each tier, projects are ranked by builder karma.
For full pricing, pack details, and a comparison table, see the Pricing page →
Jellar uses one-time review packs - not subscriptions. You buy a pack, your project goes live for the listing window, and reviews come in as testers complete missions. When the window or review cap is hit, the listing archives. Buy another pack to relist.
Sponsored projects sit above all packs at the top of the Ladder in gold. Within each band, projects are sorted by builder karma - the more karma you have, the higher you appear.
Reviews per pack
Each pack has a review cap - the maximum number of testers who can review your project in that listing window. Once the cap is reached, no new testers can start a review. See the Pricing page for exact numbers per pack.
Active domains
Free listings allow 1 active domain. Paid packs allow multiple simultaneous listings across different domains. Refer to the Pricing page for specifics.
Listing duration
How long your project stays live on the Pulse before it automatically expires. Expired listings are archived - testers can no longer start new reviews. Expiry keeps the Pulse fresh and ensures testers see current builds, not abandoned ones. Duration varies by pack - see Pricing.
Test before you list (free path only)
To keep Jellar reciprocal, free listings require you to complete 3 missions per listing slot as a tester before listing your own project. Paid packs skip this requirement entirely - you can list immediately. This ensures builders understand what it feels like to be a tester, which makes their mission briefs better.
Ladder position
The Pulse Ladder is divided into bands by pack type. Sponsored projects sit at the top in gold. Launch Pack below that, then Signal, Starter, and Free at the base. Within each band, projects are sorted by builder karma - the more karma you have, the higher you appear. Higher position = more visibility = more testers finding your project.
AI pattern summary
Signal and Launch pack buyers receive an AI-assisted analysis across all their feedback sessions. This surfaces patterns - common friction points, contradictions between testers, what most people struggled with, what most people loved - and produces a prioritised list of recommended fixes. Coming soon.
Sponsored slot
A sponsored slot places your project at the very top of the Ladder in gold for 7 days, above all pack bands. Your globe marker also appears in gold. Testers earn double karma (2× credits) for completing sponsored missions - this pulls testers to your mission first. Limited slots available. See Pricing for details.
What's in a review?
Every Jellar review follows the same structured format. No blank-page feedback. No "looks nice!" answers. Here's what every review contains.
Structured feedback format
Every tester completes the same set of fields - so you can compare reviews apples-to-apples. The format is designed to extract actionable signal, not vague opinions. All fields are enforced: testers can't skip friction points or submit one-word answers.
Friction points
Where did the tester get stuck, confused, or want to give up? Testers describe the specific moment - what they were trying to do and what went wrong. Friction points are the most actionable part of a review: they tell you exactly what to fix. Testers can add multiple friction points per review.
Example: "Sign-up button wasn't visible on mobile - had to scroll past three sections. Nearly gave up."
Magic moments
What worked better than expected? What made the tester think "oh, that's clever"? Magic moments tell you what to keep, double down on, and lead with in your marketing. They're optional but heavily weighted in karma scoring - a tester who spots something good is more valuable.
Example: "The onboarding checklist was genuinely satisfying to complete. Kept me going."
If I could change one thing
Testers are forced to prioritise. They can only name one thing they'd change. This is the most signal-dense field in the review - it tells you what a real user considers the single biggest problem. When multiple testers name the same thing, that's your top priority fix.
Would you pay for this?
A key conversion signal. Testers choose from: "Yes - without hesitation", "Probably - if the price was right", "Only if free", or "No - not for me". Available on paid-tier reviews. Across multiple reviews, this tells you whether your value proposition is landing.
Would you use this again?
Retention signal. After actually using the product, would this tester come back? Combined with "Would you pay", these two fields give you a clear picture of whether users are finding real value - or just completing the mission and moving on.
Screenshot evidence (Launch pack only)
On Launch pack projects, testers can attach up to 3 screenshots to their review - showing exactly what they saw on screen. A confusing layout, a broken button, a moment that delighted them: instead of describing it, they can show it. This turns written feedback into visual evidence you can act on immediately.
Screenshots are reviewed by admin alongside the rest of the feedback and only become visible to you once the review is approved. This is an exclusive Launch pack benefit - the highest-signal feedback for serious launch prep.
Review quality and karma weighting
Not all reviews are equal. Jellar scores reviews for quality - detail, specificity, minimum length, and non-repetition. High-karma testers (Trusted ✓ and Elite ✓) have their reviews weighted more heavily in your Signal Score. Thin or spam-detected reviews are flagged and may be voided by admin before reaching your dashboard.
Tester tiers: New (0–29 karma) → Active (30–79) → Trusted ✓ (80–199) → Elite ✓ (200+)
Pulse Boost
A one-time $15 add-on that pushes your project to the top of its Ladder band for 72 hours.
What Pulse Boost does
Every project sits in a band on the Ladder based on its pack tier. Within that band, position is determined by builder karma. Pulse Boost temporarily overrides this - your project floats to the top of its band regardless of karma, for 72 hours. After the boost window, your project returns to its normal karma-based position.
When to use it
Boost is most effective when: traffic is low and your project has stalled mid-listing window, you've shipped a significant update and want fresh eyes, or your listing is close to expiry and you want a final push before it archives.
Who can use Pulse Boost
Pulse Boost is available to paid pack holders (Starter, Signal, Launch) only. Free path listings are not eligible - the free path is designed for reciprocal testing, not promotion. To unlock Boost, upgrade to any paid pack.
Real reviews. Real exposure.
Jellar isn't just a feedback tool - submitting your project creates permanent, indexed public visibility for your build.
Your project stays visible
Every project submitted to Jellar gets a public-facing profile page linked from the builder's profile. Even after a listing expires, the project URL and title remain visible on your profile page - permanently. This means your build gets an ongoing indexed link from jellar.io, which counts as a real backlink for search engine purposes.
Leaderboard visibility
Top testers appear on the Jellar Leaderboard - a publicly indexed page. Builders with high karma also appear on the live Pulse Ladder. Both pages are crawlable by search engines, creating additional indexed mentions of your project and profile.
Signal Pack & Launch Pack: Real exposure, not just reviews
Signal and Launch packs include your project on the mid and top tiers of the Pulse Ladder - the most visible positions on the platform. More testers see your project first. Higher tester volume means more feedback AND more people encountering your URL in a trusted, indexed context. Real reviews. Real exposure.
Signal Pack: 15 reviews · 21-day listing · Mid-Ladder position · AI pattern summary
Launch Pack: 30 reviews · 30-day listing · Top non-sponsored position · AI summary + priority
Pre-flight checks
These run automatically in your browser when you hit submit. Here's exactly what each one checks.
01 - HTTPS required
Your URL must start with https://. Unencrypted http:// sites are a risk to testers and will not be accepted.
02 - Valid URL format
Must be a properly formed URL with a real domain. Malformed entries, single words, or incomplete URLs fail immediately.
03 - Not a raw IP address
Blocks URLs that point directly to IP addresses (e.g. https://192.168.1.1). Testers need a real domain, not a server address.
04 - Not a URL shortener
Blocks bit.ly, tinyurl.com, t.co, goo.gl, ow.ly and similar. Shorteners hide the real destination from testers.
05 - Allowed TLD
Rejects .tk .ml .ga .cf .gq .zip .mov and other TLDs disproportionately used for abuse and phishing.
06 - No phishing patterns
Scans the URL for brand impersonation patterns: paypal-secure, apple-login, google-verify, suspicious prefixes like signin-, verify-, etc.
07 - No lookalike characters
Rejects punycode (xn--) and non-ASCII characters used to impersonate real domains (e.g. Cyrillic letters that look identical to English ones).
08 - URL content allowed
Checks whether the URL resolves to content that is permissible on Jellar - not a raw file download, blocked service, or restricted content type.
09 - Input quality check
Your submission answers (title, description, mission brief, target user) are checked for gibberish, keyboard mashing, excessive repetition, and spam patterns. Thin or incoherent submissions are rejected.
10 - Account at least 5 days old
Your Jellar account must be at least 5 days old before you can list a project. This prevents throwaway account abuse.
11 - URL not already submitted
Blocks duplicate submissions of the same domain. If your URL is already listed and you're the legitimate owner, contact us.
12 - Site appears online and reachable
Confirms your site is actually live and accessible. A fail here means the site is down, not deployed yet, or is blocking external requests.
12 - Content matches description
A keyword-based check between your submitted description and the site's visible text content. Catches obvious mismatches - e.g. you described an invoicing tool but the site is a recipe blog.
Manual review
After automated checks pass, a human admin opens your site and checks:
- Does the site match what you described in the Vibe Check?
- Is the mission actually doable in the stated time?
- Are there hidden dark patterns, misleading flows, or tester risks?
- If test credentials were provided, do they work?
- If AI is claimed, is it actually present?
Most submissions are reviewed within 24 hours. You'll see status updates on your profile page.
Doing a mission
- Pick a project from the Pulse Ladder on the homepage.
- Read the Mission Briefing - it tells you what the builder wants you to do, what the project is about, and any warnings (login required, data collection, etc.).
- Click Open Project + Start Timer. This opens the project in a new tab and starts a 3-minute minimum dwell timer.
- Use the project as a real user would. Actually try to complete the mission.
- Come back to Jellar and fill in the feedback form. The more you fill in, the more credits you earn.
- Submit. Credits and karma are awarded immediately.
You can only review each project once. If the builder submits a new version (new URL or updated Vibe Check), you can review again.
Testers on Signal or Launch tier unlock two extra fields when reviewing any project. These questions collect higher-signal data for builders and earn the tester additional credits.
The moment the page loaded - what was the gut reaction? Within the first 5 seconds, what did you understand (or not) about what the product does? This is one of the highest-value data points for a builder: first impressions are never recoverable, so they need a real person's honest account.
Honest conversion check - not based on price, based on perceived value. Options range from "Yes, gladly" to "No - only if it were free." The tester can also add a sentence explaining what drives that answer. This tells builders where they sit on the value perception spectrum.
These fields only appear for testers on a paid tier. Builders on any tier benefit from paid-tier tester feedback - the tier is the tester's, not the builder's.
Karma & Credits
There are two separate karma scores:
Tester Karma
Earned by completing reviews. Higher tester karma puts you on the Top Testers leaderboard. Leaderboard position boosts visibility - especially if you have a live project, since clicking your name highlights your build.
🔥 Sponsored missions pay DOUBLE karma - complete a sponsored mission to earn 2× credits. Look for the gold badge on the Ladder.
Builder Karma
Earned when testers complete your missions. Higher builder karma moves your project up the Ladder within your tier. More visibility = more reviews = more karma = more visibility.
Every new account starts with 50 karma as a welcome bonus. You begin in the ACTIVE tier - not New - from the moment you sign up.
Karma is only awarded for admin-approved reviews. Every review goes into a queue before the builder sees it. Rejected or voided reviews have their karma reversed. Write like the builder is relying on you - because they are.
Rewarding great reviews
When you receive reviews on your project, you'll see a +1 karma button next to each one on your profile page. If a tester gave you genuinely useful, specific feedback - use it. It's a small signal that says "this person did the work."
Go to your Profile page → Reviews Received. Each review card has a "+1 karma" button. Click it to award one karma point to that tester. You can only award it once per review.
Karma affects a tester's position in the leaderboard and their reputation on the platform. Testers with higher karma are more visible to other builders. Quality feedback gets rewarded.
Karma earns you a reputation tier. Builders can see your tier on every review you submit - higher-tier reviewers carry more weight and trust.
All new accounts start with 50 karma, so you begin in ACTIVE tier from day one - not New.
Trusted and Elite reviewers show a weighted badge on their reviews. Builders see this when reading feedback - it signals a track record of quality. The only way to climb is to give consistently good feedback over time.
To keep the leaderboard and Ladder reflecting active community members, karma decays for inactive accounts.
- If you have not submitted any reviews and have not signed in for 30 days, your karma drops by 5 points flat on the 1st of the following month.
- Decay is floored at 0 - you will not go negative.
- You will receive an email warning 7 days before decay kicks in so you have time to complete a review and reset the clock.
The fix is simple: complete one review, and the 30-day clock resets.
Feedback quality
Good feedback is specific. "The UX is bad" is useless to a builder. "The sign-up button wasn't visible on mobile - I had to scroll down to find it" is gold.
Good feedback
- Specific friction points with context
- What you expected vs what happened
- Your honest "would I use this again" answer
- The single most important thing to fix
What gets flagged
- Vague one-liners with no detail
- Copy-pasted responses across sessions
- Profanity or abusive language
- Submitting feedback without actually using the product
What gets you banned
- Submitting a project you don't own without permission
- Lying on the Vibe Check questionnaire (hiding data collection, payments, or login requirements)
- Submitting phishing, malware, scam, or illegal content
- Using Jellar to harvest test credentials or personal info from testers
- Creating throwaway accounts to bypass submission limits or review gates
- Submitting fake or bot-generated feedback
Bans are tied to your GitHub account and IP. Ban appeals can be sent to the email on your GitHub profile.
Language policy
All project submissions, mission briefs, and feedback must currently be written in English. This is a platform requirement during early access to ensure our team can review all content. Support for additional languages is on the roadmap - see the Roadmap for details.
Jellar is a professional feedback platform. The system automatically blocks profanity and prohibited language from all submissions and feedback. If your text contains profanity, the field will be highlighted and you'll be asked to remove it before you can proceed. It simply won't go through.
This applies to project titles, mission briefs, descriptions, and all tester feedback fields.
FAQ
Still have questions? Reach out via GitHub or email.
READY TO SUBMIT →About the name
The name JELLAR traces back to Gjallar - a word from Old Norse meaning to resound, to yell, to broadcast a signal. In Norse mythology, the Gjallarhorn is the horn of Heimdall, the watchman of the gods. When sounded, it sends a signal across all nine worlds - a call to attention that cannot be ignored.
That's the idea behind this platform. Builders ship things into the void. Jellar is the horn - it broadcasts their signal to the humans who will actually listen, test it, and send back a real response.
The original team behind Jellar has Scandinavian heritage, and Norse mythology felt like the right well to draw from - timeless, a little dramatic, and genuinely about communication across distance. It also has a stellar ring to it, which felt right for something meant to travel far.
The domain was available. The gods aligned. Here we are.