DEV Community

Luca Tomasino
Luca Tomasino

Posted on

Stop undercharging: I built an AI pricing co-pilot in MeDo

#ai #buildinpublic #builtwithmedo

Most freelancers I know lose money on every project. Not because they're bad at their craft, but because they're bad at pricing.

I see it all the time. A one-line quote, no rationale, a single number that gets negotiated down to nothing. Independent professionals consistently undercharge by 30 to 50 percent versus market rates. I'm a freelance AI automation consultant in Italy and I send quotes every week. It's still painful, every time.

So when I saw the Build with MeDo Hackathon, I built the tool I wished I had when I started.

It's called FairQuote. This post is the build-in-public log of v0.1.

The idea

FairQuote is an AI pricing co-pilot. You describe a client project in a chat. The AI interviews you adaptively, asking 6 to 10 smart questions about scope, deliverables, urgency, and your context. Then it generates a proposal with three pricing tiers (Lean, Standard, Premium), each with deliverables, milestones, payment schedule, and explicit pricing rationale.

The differentiator versus the dozen quote calculators out there:

  1. Multi-turn intake that adapts to project type. A designer gets different questions than a developer.
  2. Three-tier anchoring output. Research on negotiation shows presenting Good/Better/Best increases close rates and average deal size.
  3. Pricing rationale built in. Every number comes with hourly equivalent, market benchmark, and value justification you can paste straight into negotiation emails.

Why MeDo

I'm not a frontend developer. I build automations and AI agents in n8n, I'm comfortable with backend logic and prompt engineering, but spinning up a full-stack web app with auth, dashboards, and document generation is not where I want to spend my hackathon week.

MeDo is Baidu's no-code agentic app builder. You describe what you want in natural language, MeDo scaffolds the frontend, backend, database, and integrations. It plugs into ERNIE for LLM features. Deploy is one click.

Perfect for a hackathon constraint of one week and limited time budget.

The master prompt

This was the first prompt I gave MeDo. One shot, no iteration yet.

Build a single-page interactive demo of FairQuote, an AI-powered
pricing co-pilot for freelancers and independent professionals.

The page has three sections, top to bottom:

1. LANDING (hero): headline, subhead, CTA "Try the demo" that
   scrolls to the chat below.

2. AI CHAT (interview): a chat interface where the AI assistant
   (Baidu LLM Plugin / ERNIE) interviews the user about a
   hypothetical client project. One question at a time. Adaptive
   follow-ups. Cover: project type, deliverables, timeline,
   hourly target, country.

3. PROPOSAL OUTPUT: three cards side by side (Lean / Standard /
   Premium). Each card shows tier name, total price, 3 to 5
   deliverables, and a short rationale paragraph.

DESIGN
Dark mode. Clean sans-serif. Deep teal accent. No flashy gradients.
Enter fullscreen mode Exit fullscreen mode

What MeDo produced from that single prompt: a working landing page, a chat interface wired to ERNIE, and a three-tier comparison view. Not perfect, but shipped and deployed in minutes.

The three-tier logic

Lean covers exactly what the client asked for, no extras, longer timeline, fewer revisions. Standard is the suggested choice with light extras (testing, two revisions, basic documentation). Premium is the value-driven upsell (post-launch support, analytics, training, faster turnaround).

Pricing scales not linearly but on perceived value. Lean is the anchor low, Standard is priced at the freelancer's real target, Premium sits at 1.6 to 2x Standard. Most clients self-select into the middle, which is exactly the point.

Every tier text generated by ERNIE includes three things:

  • Hourly equivalent at the freelancer's stated target
  • Market benchmark for similar scope
  • Value justification tied to the client's stated needs

This is what makes the rationale defensible in negotiation, not just numbers on a card.

Demo

Live app (v0.1): https://app-b85fruavsxkx.appmedo.com

It's a barebones v0.1. The chat works, the three tiers generate, the layout is clean. PDF export, persistent storage, and a quote management dashboard are coming in v1 by May 20.


What I learned about MeDo so far

The good:

  • One master prompt to a working full-stack app is real, not marketing fluff.
  • ERNIE plugin wires in cleanly. No agent orchestration code on my side.
  • The visual editor catches small UI tweaks faster than re-prompting.
  • Deploy is genuinely one click.

The friction:

  • Some prompt edits regenerate parts of the app you didn't want touched. Save versions before big iterations.
  • Rate limits kick in during heavy testing. Plan around it.
  • The chat-to-state transitions need explicit instructions or they get fuzzy.

Net assessment: for an idea that is 70% AI behavior and 30% UI, MeDo is a serious accelerator. I would not have shipped a v0.1 in this timebox without it.

What's next

Before May 20 I want to ship:

  • User accounts and persistent quote history
  • PDF export with a clean professional template
  • Cover email generator (copy-to-clipboard ready)
  • A few canned templates per profession (designer, developer, copywriter, consultant)

If you're a freelancer who's tired of quote anxiety, try the demo and tell me what's missing. I'm building this in public for the next few weeks.

Built with MeDo by @Medo_CodeFree for the Build with MeDo Hackathon.

BuiltWithMeDo

Top comments (0)