The base44 developer vetting guide.

1. Why base44 hiring is hard

Base44 is a fast-moving AI vibe-coding platform. Wix acquired it for $80M in June 2025 and the platform has shipped meaningful behaviour changes every quarter since. Three structural problems make hiring on it harder than hiring for traditional stacks.

No certification body.Base44 has a partner program, but it is invisible to most buyers, does not publish a skills assessment, and does not function as a meaningful credential. There is no "Certified Base44 Developer" equivalent to AWS Certified Solutions Architect. Anyone can claim expertise.

Tribal platform knowledge. Documented base44 behaviours — function routing breaking after deploy, credit-burn loops on trivial edits, SSO bypass via header manipulation, Stripe webhook regressions after platform updates, AI-agent regression loops — are tribal. They live in the feedback portal, in private Slack channels, and in the heads of engineers who have shipped on the platform. A generalist who learned base44 last quarter will not have seen them.

Marketplace incentives are misaligned. Fiverr rewards low headline rates (race to the bottom). Upwork rewards profile gaming and review collection. Generalist agencies reward bench utilisation, not platform expertise. None of these are evil — they are just not selecting for what you actually need on a base44 engagement.

The framework below corrects for all three. It is the same framework we use to vet our own engineers, published so you can apply it to any candidate.

2. The five-gate framework

Five gates. Each is independent — a candidate must pass all five. Failing any single gate is sufficient cause to reject.

1. Gate 01

   Reps

   Minimum ten base44 apps shipped to production. Not "tutorials completed" or "test apps in the editor." Production apps with live URLs, real users, and at least one platform update survived. Ask for named portfolio links and verify they load.

2. Gate 02

   Pattern coverage

   Documented playbook for every error type in the base44 feedback portal. A real specialist names function routing, credit burn, SSO bypass, Stripe webhook breakage, and AI-agent regression loops in under a minute when prompted. A generalist hedges or asks for the documentation.

3. Gate 03

   Code review

   Every fix is peer-reviewed by a second senior engineer before ship. Sprint output is never single-authored on critical paths (auth, payments, data). Single-engineer freelancers cannot pass this gate; insist on a peer review process or a second pair of eyes.

4. Gate 04

   Written communication

   Every engagement closes with a written summary: what changed, why, what to monitor. Engineers who cannot write a clean three-paragraph technical summary do not pass. Verbal-only handoff is an immediate disqualification on production work.

5. Gate 05

   Bylined accountability

   Named engineer on every engagement letter. LinkedIn or equivalent professional profile. The engineer who scopes the work is the engineer who ships the work. Bench-and-rotate (engineer-A scopes, undisclosed engineer-B ships) fails this gate.

3. Portfolio review checklist

Spend thirty minutes on every shortlisted candidate's portfolio. Look for these eight signals.

• · Live URLs. The portfolio should include URLs you can load. Screenshots are insufficient — anyone can ship a screenshot.
• · Production traffic indicators. Apps with real users (visible reviews, app-store presence, social proof). A portfolio of internal tools no one else uses is harder to verify.
• · Diversity of use case. Look for at least three different verticals: SaaS, marketplace, internal tool, e-commerce. A portfolio of ten near-identical CRUD apps is a junior tell.
• · Auth and payments. Apps with non-trivial auth flows (SSO, OAuth, multi-tenant) and live payments (Stripe, PayPal). Both are where base44 expertise actually pays off.
• · Date stamps. Recent shipped work (in the last six months). Base44 changes fast; portfolio work from 2024 may not reflect current platform behaviour.
• · Named clients (where consent exists). A few named clients with permission to share. All-anonymised portfolios are not a red flag in themselves but reduce your verification options.
• · Public case studies or write-ups. A senior engineer typically has at least one public technical write-up — a blog post, a Stack Overflow answer, a GitHub repo. Zero public footprint is a junior tell.
• · Migration or rescue work. Bonus: candidates who have rescued a broken base44 app or migrated one off the platform have seen the platform's edges.

4. The twelve technical screening questions

Run these in a thirty-minute screening call. A real specialist answers in under a minute per question with specific platform terminology. A generalist hedges, asks for documentation, or talks about no-code platforms in general.

1. How would you diagnose a base44 backend function returning 404 in production but working in preview?
2. What is the most common cause of credit-burn loops on trivial edits, and how do you stop them?
3. Walk through the SSO bypass via header manipulation. How would you mitigate it on a live app?
4. Describe how Stripe webhook handlers can break after a base44 platform update and how you restore them.
5. How does base44's client-side rendering affect Google indexing, and what is your usual workaround?
6. What causes data-loss when users return to an app, and what is your usual remediation?
7. What rate-limit thresholds have you seen base44 enforce in production, and how do you handle 429 backoff?
8. When would you advise a client to migrate off base44 rather than fix in place?
9. What is your peer-review process on critical-path code (auth, payments, data)?
10. Walk through how you would scope a $1,500 single-bug fix-sprint engagement end-to-end.
11. How do you document a fix so it survives the next platform update?
12. What does your written hand-off deliverable look like? Can you share a redacted example?

Score each question 0-2: 0 = no specific answer, 1 = generic answer, 2 = specific platform-aware answer with concrete example. A passing score is 18+ out of 24. Below 12, reject. 12-17, run a paid trial task before committing.

5. The paid trial task

For any engagement over $2,000, run a paid two-hour trial task before committing to the full scope. The trial separates pattern-matchers from generalists more reliably than any conversation.

The brief. Pick a documented base44 issue from the feedback portal — function routing returning 404, credit burn on trivial edits, SSO bypass — and ask the candidate to: (a) describe the root cause in one paragraph, (b) outline a remediation plan in three to five steps, (c) write a regression test plan. Do not ask them to ship code; the trial is about diagnostic and communication quality.

What to score.Specificity (do they name platform behaviours, or talk in generalities?). Causal clarity (do they trace the bug from symptom to root cause, or list possibilities?). Written quality (would you hand this to a board member?). Time discipline (did they deliver inside two hours?). Cost: pay $200-$500 for a senior's two hours.

Never ask for free work. Free-trial requests are an industry red flag and the best engineers will refuse. A paid trial signals that you are a serious buyer and respects their time.

6. The six red flags

• Unwilling to share named portfolio. If they cannot show shipped work in their own name, they probably did not ship it. Walk away.
• Hourly billing with no scope cap. Open-ended hourly is a license to drift. Insist on a fixed scope or a hard cap. "Hourly with weekly check-ins" is not a cap.
• No written deliverable. Verbal-only handoff is unmaintainable. The engagement letter must specify a written summary deliverable.
• Bench-and-rotate. Engineer-A scopes the work, engineer-B (often offshore, often undisclosed) ships it. Common at large generalist agencies. Insist on the named engineer end-to-end.
• Subcontracting offshore without disclosure. Common on Upwork and Fiverr. Ask explicitly: "Will any other engineer touch this work? If yes, who, and what is their LinkedIn?" Get the answer in writing.
• Reviews that all sound the same. Marketplace reviews are often coordinated. Look for specific project detail (not generic five-star copy-paste). Three reviews with detail beat thirty without.

7. Rate benchmarks

Use these as a sanity check. A rate dramatically below the band typically means the candidate is junior, subcontracting, or undercutting on quality. A rate dramatically above means they are positioning for an open-ended retainer rather than a defined scope.

• · Fiverr / entry freelance: $10-$50/hour. Cosmetic work only.
• · Upwork / mid-tier freelance: $25-$100/hour. Variable quality.
• · Offshore generalist agency: $22-$45/hour, $2,880-$6,000/month retainer.
• · US/EU generalist agency: $80-$150/hour. Account-manager layer.
• · Boutique tech-hub agency: $150-$250/hour. Strong engineering culture.
• · Specialist team (fixed-scope): $497 audit, $1,500 sprint, $3,000 multi-bug, $4,500-$15,000 builds, $6,000-$25,000+ migrations.
• · Specialist hourly (legal/DD only): $200-$400/hour, four-hour minimum.

See our full rates page for the source data and hidden-cost analysis.

8. Reference checks

Spend thirty minutes on references. Two questions matter more than the rest:

1. What did the engineer do when something went wrong? Every engagement has a wobble. The honest reference describes how the candidate handled it. The dishonest reference says "nothing ever went wrong" — which is a tell.
2. Would you hire them again, today, at the same rate? Not "were you happy." The willingness-to-rehire test eliminates polite recommendations.

A candidate who cannot produce two references with verifiable LinkedIn profiles is not a candidate. Walk away.

9. Engagement letter checklist

Before any payment, the engagement letter must specify all of the following. If any are missing, the engagement is structurally weak even if the engineer is strong.

• · Named engineer (or named engineers, if peer-reviewed).
• · Fixed scope or capped hourly with a written cap.
• · Written deliverables: code, summary, regression plan.
• · Timeline with milestones.
• · Mutual NDA terms.
• · Money-back or rework clause if the work fails.
• · Subcontracting clause: explicitly named, or explicitly forbidden.
• · Post-ship support window (usually 14 days).
• · IP ownership clause.
• · Termination terms.

10. Apply this framework, even to us

We publish this framework because we want it applied to our own engineers. If you are evaluating Base44Devs for an engagement, run us through every gate, every red flag, every reference check. Ask the twelve technical questions on the screening call. Run a paid trial task on a $497 audit (it doubles as the trial) and use the result to decide whether to commit to a sprint, build, or migration.

If you have shortlisted two or three base44 specialists and need an independent technical screen before committing, we offer a $497 candidate-screening engagement: we run a candidate against the framework, conduct the technical screening call, and deliver a written verdict. Useful when you want a second opinion that has no incentive to either flatter or undermine the candidate.