Lemonsignal
AI skills assessment: candidates do a real task in Claude Code or ChatGPT and you see every prompt.
From $299/20 candidates Free trial
Visit Lemonsignal (opens in a new tab) Follow Lemonsignal updates
Bottom line
AI skills assessment for hiring: candidates do a real task in Claude Code, Claude Cowork, ChatGPT or Codex in the browser, and you get the finished work plus every prompt, command and edit as a session replay. Each candidate gets a role score and six AI skill ratings, each linked to the moment it is based on, and a question to ask in the interview. It is free on 3 candidates; packages start at $299 for 20 candidates and monthly plans at $449.
Stands out for: It tests how someone works with AI tools instead of banning them. Every score links to a timestamp in the session, so a hiring manager can watch the moment a candidate caught an AI edit nobody asked for, and the result ends with an interview question about the weakest moment. Tasks cover engineers, marketers and analysts, or you bring a real piece of work from your team.
Watch out for: Pricing is per candidate who hands in the task, so screening at volume gets expensive ($10 per extra candidate on monthly plans). Getting started means a 30-minute setup call rather than self-serve signup, unused monthly candidates do not carry over, and it is a new product with no outside reviews yet.
Overview
Lemonsignal is an AI skills assessment built around one question: how does this candidate actually work with AI? Instead of a quiz, candidates do a real task in the AI tools they would use on the job, and the hiring team sees all of it.
You pick a task from Lemonsignal's library for engineers, marketers, analysts and other roles, or bring a real piece of work from your team, and send the candidate a link from your dashboard. The candidate works in Claude Code, Claude Cowork, ChatGPT or Codex in the browser, with nothing to install, which the vendor says keeps security risk down.
When they hand in, you get the finished work and a session replay of every prompt, command and file edit, alongside a readable timeline. Each candidate gets an overall AI skills score for the role and a rating on six skills: understanding the task before handing it to AI, briefing the AI, checking its output, correcting course, signing off work that is ready to send, and making the work repeatable. Every rating links to the moment it is based on.
The result ends with a suggested interview question about a specific moment, such as why the candidate did not tell the AI which files to leave alone. The first role is set up with you on a 30-minute call.
You can try it free on 3 candidates. After that, packages of 20 or 50 candidates cost $299 or $599 and last 12 months, and monthly plans run $449, $799 or $1,399 for 50, 100 or 200 candidates, with extras at $10 each and 2 months free on yearly billing. Only candidates who hand in the task are charged.
Who it's for
Best for
- Teams hiring engineers who will code with AI assistants
- Marketing and analyst roles where AI fluency matters
- Hiring managers who want evidence of AI skill, not a self-rating
Not ideal for
- High volume screening on a tight budget
- Roles where AI tools are not used on the job
Key features
Real tasks with real AI tools. Pick a task from the library for engineers, marketers, analysts and more, or bring your own. Candidates do it in Claude Code, Claude Cowork, ChatGPT or Codex in the browser, with nothing to install.
Session replay. Every prompt, command and edit lands in your dashboard, with a readable timeline of what happened.
Six AI skill scores. Understand, Brief, Check, Correct, Sign off and Reuse, each linked to the moment in the session it is based on, plus an overall score for the role.
Interview question. Each result ends with a question to ask the candidate about a specific moment.
AI features
| What the AI does | Plan |
|---|---|
| Candidates work with Claude Code, Cowork, ChatGPT or Codex inside the assessment | All plans |
Only features named in the vendor's own documentation. Anything sold separately is marked as an add-on.
Pricing
| Plan | Price | What you get |
|---|---|---|
| 20 candidates | $299 | Paid once, use within 12 months |
| 50 candidates | $599 | Paid once, use within 12 months |
| 50 a month | $449/mo | Extra candidates $10 each |
| 100 a month | $799/mo | Extra candidates $10 each |
| 200 a month | $1,399/mo | Extra candidates $10 each |
Try it free on 3 candidates, no card. Packages are paid once and used within 12 months: $299 for 20 candidates or $599 for 50. Monthly plans are $449/mo for 50 candidates, $799/mo for 100 and $1,399/mo for 200, with extra candidates at $10 each; paying yearly gives 2 months free. A candidate counts only once they hand in the task. Confirm current terms on the vendor site.
Checked 6 Nov 2026. Vendors change prices without notice.
Is Lemonsignal expensive?
Lemonsignal starts at $299/20 candidates. We list 70 assessments tools in total. 13 have a free tier, and those with a public price typically start around $80 a month.
- 36 quote-based
- 13 free tier
- 37 public price
Worked out automatically from every assessments listing, so it updates as the directory changes.
Pros and cons
Pros
- Candidates use real AI tools in the browser with nothing to install
- Full session replay with every prompt
- Only pay for candidates who hand in the task
- Free on 3 candidates, no card
Cons
- Per candidate pricing adds up at volume
- Setup call required to start
- New product with no outside reviews yet
Integrations
Claude Code Claude Cowork ChatGPT Codex
What's new at Lemonsignal
No updates logged yet. Follow Lemonsignal and we will email you when it announces something new or its starting price changes.
Follow Lemonsignal
Get an email when Lemonsignal ships a new feature, posts release notes or changes its starting price. Only when something changes, and you can stop in one click.
How Lemonsignal compares
See all Lemonsignal alternatives, with what each one does differently and a head to head for every pairing.
| Tool | Free tier | From | MCP | API | Compare |
|---|---|---|---|---|---|
| | – | $299/20 candidates | – | – | |
| | ✓ | $215/mo | – | – | Compare |
| | – | $79/mo | ✓ | ✓ | Compare |
| | – | $75/mo | – | – | Compare |
$215/mo Free tier
Broad skills test library with a free plan and paid plans from $215/mo, for screening many candidates on general skills rather than AI fluency.
$79/mo MCP API
Coding assessments from $79/mo in its own IDE, if classic technical screens matter more than evidence of AI workflow.
$75/mo
Job simulation assessments from $75/mo, a cheaper route to realistic tasks without the AI session replay.
Your next step
Shortlisting Lemonsignal?
Still deciding? Ask an AI assistant about Lemonsignal.
ChatGPT (opens in a new tab)Claude (opens in a new tab)Grok (opens in a new tab)
Frequently asked questions
What is Lemonsignal?
Is Lemonsignal free?
Does Lemonsignal have an MCP server or API?
What are alternatives to Lemonsignal?
Is Lemonsignal your product? Add the Listed on Work Insiders badge
Put this on your site so buyers can read an independent write up. It links to this page.
More about the badge: workinsiders.com/badge. Something wrong in this listing? Tell us.