Question 1

Which model do you ship on?

Accepted Answer

Default is Claude Sonnet 4.6 with prompt caching — best reasoning-to-cost ratio for most workloads. We switch to Claude Opus 4.7 for high-stakes reasoning, OpenAI GPT-5/4o where its function-calling, vision, or Realtime voice API is the better fit, and Gemini Flash for high-volume cheap inference. Provider-agnostic from day 1: a thin abstraction over the model API so we can swap Claude → GPT → Gemini in a config flip.

Question 2

What about hallucinations?

Accepted Answer

Three layers. (1) RAG against your data instead of relying on model knowledge — the model retrieves your actual records, contracts, or docs and answers from them. (2) Function-calling for structured output — the model returns JSON that matches a schema, not free-form text. (3) Human-in-the-loop on high-risk paths — the model proposes, your team approves before any irreversible action. We don't ship AI on actions you can't undo.

Question 3

Do you fine-tune?

Accepted Answer

Rarely. Prompt engineering plus retrieval gets 90% of the value at 5% of the cost — and it stays portable across providers. Fine-tuning locks you to one model and one snapshot, and you re-do the work every time the base model upgrades. We fine-tune only when latency or cost forces it, and even then we'd rather distill into a smaller open model than fine-tune a closed one.

Question 4

What if scope changes mid-project?

Accepted Answer

Small additions are absorbed in the timeline buffer we build into every quote. Anything that grows scope by more than 20% triggers a quick re-quote — written, signed, and gated by an explicit change-order email before any extra work starts. No surprise invoices.

Question 5

Why not hire an agency for $50K?

Accepted Answer

Agency overhead typically runs 40-50%. Half your money pays for an account manager, a project manager, and the meetings between them. The actual code gets written by a junior. With Wolrix, you pay the technical lead doing the work — directly.

Question 6

How does the satisfaction guarantee work?

Accepted Answer

If you're not satisfied with the deliverable measured against the agreed SOW, full money back. No questions, no escalation, no haggling. Same policy we run on Upwork — verifiable on the public profile (100% Job Success Score across 42 projects).

Question 7

What happens after the 30-day post-launch support?

Accepted Answer

Two options: hourly support at $100-$200/hr for ad-hoc fixes and tweaks, or a monthly retainer for ongoing development. Most clients pick the retainer once they see the velocity. We'll quote it based on your expected workload.

Question 8

Who owns the IP?

Accepted Answer

You do. 100% of the source code transfers to you on final payment. We hand over the open repository, environment config, and deployment access. No licensing, no royalties, no strings.

Question 9

What's NOT included?

Accepted Answer

Hosting and cloud bills (Vercel, AWS, etc.), third-party API costs (Stripe, Claude API, OpenAI, Twilio, and so on), domain registration, and ongoing content or marketing work. We'll estimate these for you up front so there are no surprises — but they go on your accounts, not ours.

Question 10

Payment terms?

Accepted Answer

50% to start, 25% at the agreed mid-project milestone, 25% on delivery. Stripe or wire. Invoiced in USD. Terms are written into the SOW before kickoff.

Question 11

Do you sign our NDA?

Accepted Answer

Yes — that's the default. We'll sign yours, or bring our own if you don't have one ready. NDA goes out before the first technical conversation.

Question 12

What stack do you build on?

Accepted Answer

Default is Next.js 15 + TypeScript + Postgres + Drizzle + Stripe + Resend + Vercel + Sentry, with Claude API and OpenAI API for AI work. Battle-tested, fast to ship, easy to hand off. We deviate when there's a real reason: HIPAA constraints, existing infrastructure, specific integrations. We pick the right tool, not the trendy one. Full breakdown at /stack.

Question 13

Can we hire your team afterward?

Accepted Answer

Post-launch we offer part-time retainers — same technical lead, ongoing access. We don't transfer staff or do contract-to-hire conversions. If you need a full-time team, we'll help you scope and recruit one.