AFF Lab

Custom software for the work no off-the-shelf tool does

When you outgrow Zapier and your CRM's "advanced workflows", we build the thing properly. Sales tooling, AI workflows, scrapers, integrations — anything that needs code.

Discuss your project

Production receipts

5+
Production systems running unsupervised in client environments
100%
Codebases handed over with docs and architecture diagrams
0
Surprise tech debt incidents three months in
2–8w
Typical scope from kickoff to deployed v1
1

Production-grade from day one

We build systems we'd run ourselves. No throwaway scripts, no surprise tech debt three months in.

2

Sales and outreach DNA

We understand what you're trying to automate because we run the same operations. The brief lands faster.

3

AI where it actually helps

Not "AI everything" — AI where it earns its keep. Enrichment, classification, generation, parsing.

How a project runs

1. Scoping call

We listen to what you are trying to automate, what fails today, what success looks like. No code talk yet — just the operating problem.

2. Tech spec

Architecture, integrations, data flow, edge cases — written and signed off before code starts. No surprise re-scoping mid-build.

3. Build with weekly checkpoints

Working software demoed each week. You see progress, we catch drift early, changes are negotiated as we go.

4. Deploy + monitor

CI/CD, observability (Sentry or your stack), and an on-call playbook for the first 30 days. We watch it run.

5. Handover with docs

Code, architecture diagrams, runbook, and a knowledge-transfer session with your team. You are not dependent on us to keep it running.

What you get

  • Working software deployed to your infrastructure (or ours, if preferred)
  • Architecture documentation and sequence diagrams
  • Test suite covering critical paths
  • Monitoring and alerting wired (Sentry, Datadog, or your stack)
  • On-call playbook for common failure modes
  • Knowledge transfer session with your engineering team
  • 30 days of post-deploy iteration

Honest fit check

You will get a lot out of this if:

  • You outgrew Zapier and your CRM's "advanced workflows"
  • You need automation that compounds (sales tooling, scrapers, AI workflows)
  • You have engineers who will maintain it (or want us on retainer)
  • Project budget $10K+

Probably not us if:

  • "Build me an app" with no spec or operating problem
  • Want to replace SaaS you can just buy (Hubspot, Notion, Linear)
  • No internal owner for the system long-term
  • Need a CTO-for-hire — we are project-based, not interim leadership

The thinking behind the service

What clients usually push back on

"Why not low-code or no-code?"

Great for simple automations. Hits walls fast on edge cases, data volume, or anything custom. We build when low-code stops scaling, not before.

"In-house engineer vs you?"

In-house wins for permanent ownership. We win for fixed scope, fast delivery, no recruiting overhead. Many clients use us to build, then hand off to in-house to maintain.

"Will we be locked in?"

No. Full code ownership, all docs, clean handover. Many projects end with us out of the picture entirely.

"Why is it not cheap like Fiverr?"

Fiverr ships scripts. We ship production systems. The difference shows month three, when nothing has caught fire.

"What if requirements change mid-build?"

Built into the process. Weekly demos catch drift early. Scope changes are negotiated, not surprise invoices.

Common questions

What languages and stacks do you use?

Python (FastAPI, Django), TypeScript (Node, Bun), Astro and Next.js for frontends, PostgreSQL, Redis, Cloudflare Workers. AI: OpenAI, Anthropic, Mistral.

Where does the system run?

Your infrastructure (AWS, GCP, Azure) or ours (Hetzner, Cloudflare). Your call — we adapt to your stack.

Who owns the code?

You do. Repo lives in your GitHub from day one. We push to it; you own it.

Do you sign NDAs?

Yes, standard practice. We sign mutual NDAs before scoping calls when requested.

What about maintenance after launch?

30 days included. Retainer available ($3K–15K/month depending on system criticality and on-call expectations).

Can you work on existing codebases?

Yes. Audits, scoped refactors, modernization of legacy systems, AI integration into existing apps.

What does it cost?

Project-based. Typical range $10–50K. Quoted after scoping call once we see the actual problem.

Have a system that should just exist?

30-minute scoping call. We listen to what you are trying to automate, ask sharp questions, and tell you honestly whether code is the right answer.

Discuss your project