SaaS products
Multi-tenant platforms, auth, billing, RBAC built in.
Software that fits your business, not someone else's.
A senior product team embedded with yours, designers, PMs, engineers, all senior. Prototype in weeks, ship in months, stick around to maintain. No pass-the-parcel to juniors.
- Certified Odoo Partner
- ISO 12207 & 27001 certified delivery
- Since 2005 20 years enterprise delivery
- 50+ Odoo implementations
- 500K-employee Odoo HR & payroll deployment
Delivered from Kabul
Custom software development, delivered by senior teams based in Kabul, governed under ISO 12207 and ISO 27001. Twenty years of enterprise delivery in Afghanistan, including the 500,000-employee national-government Odoo HR rollout and Jobs.af. Engagements run from our Kabul office with regional support in Dubai and India.
- 500,000-employee Odoo HR engagement for a national government
- Jobs.af, Afghanistan's largest job-hunting platform, built by NETLINKS
- Clients include Da Afghanistan Bank, Afghanistan Payments System, AKDN, World Bank, UNDP
Why most custom software projects drift into disappointment.
Custom software fails predictably, senior architect wins the deal, hands execution to juniors, MVP ships acceptably then breaks in year one as scope outgrows the codebase. Clients pay twice.
Every squad has a senior lead on every role. Boring defaults, TypeScript, Postgres, standard observability, because boring means well-understood failure modes at 2am. Several of our builds are still in production a decade in.
What we build
Products, platforms, tools.
Internal tools
Line-of-business apps replacing spreadsheets.
Data platforms
Pipelines, warehouses, analytics tools.
Customer portals
Self-service portals reducing support load.
Chat & collaboration
Real-time messaging, doc collab, activity feeds.
Compliance apps
Audit-ready workflows for regulated industries.
How we work
Senior squads, modern defaults.
Senior lead on every role, engineer, designer, PM. Stack is boring on purpose: TypeScript/PHP, Postgres/MySQL, Kubernetes or serverless. Deploy, monitor, rollback are first-class from day one.
- Senior-only teams, no pass-through to juniors
- Stacks: TypeScript, Python, PHP, Postgres, MySQL, Kubernetes
- CI/CD, observability, IaC from day one
- ISO 12207 + ISO 27001 certified delivery
- SSO, RBAC, audit logs, encryption by default
- Design system built to extend, not locked to scope
Tech stack
Our default stack, and what we reach for beyond it.
Boring, battle-tested defaults, every item below has shipped more than once. AI-assisted tools used under the same code review and security discipline.
Languages
TypeScript 
Python 
PHP + also
Go 
Rust Web frontend
Next.js 
Astro 
React 
Vite + also
Svelte / SvelteKit 
Vue Backend / APIs
Node.js 
Hono 
NestJS 
FastAPI 
Laravel 
Livewire 
Filament + also
Django Go services Mobile
React Native + also
Swift 
Kotlin Data
PostgreSQL 
MySQL 
Redis S3-compatible storage
+ also
Clickhouse pgvector 
DynamoDB AI & ML
OpenAI 
Anthropic 
Vercel AI SDK 
LangChain pgvector + also
LlamaIndex Pinecone
Local LLMs
AI-assisted delivery
Claude Code 
Cursor 
GitHub Copilot Cloud & infrastructure
AWS Vercel 
Kubernetes 
Docker 
Terraform + also
Azure 
Google Cloud Observability
OpenTelemetry 
Sentry 
Grafana 
Prometheus CI & delivery
GitHub Actions Vercel deploys + also
GitLab CI 
CircleCI Auth & secrets
SAML / OIDC SSO
AWS Secrets Manager KMS encryption
Testing
Vitest 
Playwright 
Pytest Pest
PHPUnit
Delivered through ISO 12207 software lifecycle and ISO 27001 information security controls, held by NETLINKS Ltd.
, Honest fit
When custom is the right answer.
Build custom when
- No off-the-shelf tool covers the core workflow adequately
- The software is differentiating, it's part of your moat
- Regulatory or industry constraints require specific data handling
- You have budget for long-term maintenance (build is ~20% of TCO)
- The team to maintain it post-build exists or can be hired
Don't build custom when
- A SaaS tool can do 85% of what you need (configure, don't code)
- The requirement is likely to change quarterly (configure or buy)
- Budget doesn't cover 3-5 years of maintenance cost
- Nobody on your team can own the system after launch
- The differentiation argument is vague rather than specific
, Questions
Custom software, FAQ.
Do you do fixed-bid or time-and-materials?
Both, depending on scope clarity. MVPs are often fixed-bid after a discovery sprint. Platform work is usually T&M with sprint-level commitments. We never hide budget risk behind change-order chains.
What does a squad look like?
Typical squad: lead engineer (senior) + 2-3 engineers (mix of senior and mid) + designer (senior) + PM (senior). For larger builds, multiple squads coordinate under a single tech lead. For narrow technical tasks (a specific module or integration), smaller teams work.
What tech stack do you use?
Default stack: TypeScript, Python, and PHP as primary languages; Next.js, Astro, and React for web; Node.js (Hono, NestJS), FastAPI, and Laravel (with Livewire and Filament for admin) for backend; PostgreSQL and MySQL for transactional data; Redis for caching and queues; S3-compatible object storage; Kubernetes or serverless (Vercel, AWS Lambda) for compute; OpenTelemetry, Sentry, Grafana, and Prometheus for observability, all open source. For rapid MVPs and time-boxed engagements we use AI-assisted development (Claude Code, Cursor, GitHub Copilot) to compress cycle time. For specific domains we reach for Go on high-throughput services and Rust on performance-critical components. Full stack table on this page.
What about long-term support?
We offer support retainers covering bug fixes, enhancements, platform upgrades, and security patches. Clients commonly stay with us for 3-5 years post-launch. If you want to internalize the team, we do structured handover with 3-6 months of knowledge transfer.
Do you use offshore or nearshore?
Senior leads are staffed from the US. Engineering capacity comes from our own Gandhinagar, India delivery center and our Dubai, UAE office, all direct hires on our payroll, not subcontracted. We occasionally bring in specialist contractors (iOS Swift, niche ML, specific industry expertise) under direct supervision, never pass-the-parcel.
How do you approach testing?
Unit tests for business logic (not test-for-test's-sake). Integration tests for API contracts. End-to-end tests for critical user flows only (they're expensive to maintain). CI runs all tests on every commit. Type safety (TypeScript strict mode) catches a category of bugs before tests run.
What about security and standards?
Secure-by-default architecture: SSO via SAML/OIDC, role-based access control, encryption at rest and in transit, audit logging, secrets in a vault (not env vars). Dependency scanning in CI. Penetration testing before production for security-sensitive apps. Information security governed by our group's ISO 27001 certification; software lifecycle governed by ISO 12207. HIPAA and PCI extensions configured when required.
What about AI/ML features?
We integrate AI features (LLM-powered search, embeddings, agent workflows) where they provide real value. Rigorously, not trendily, every AI feature has clear success criteria, failure modes, observability, and cost tracking. See our agentic AI practice for detail.
Custom build coming up?
30-min architecture review with a senior architect, not an SDR. We map the build, surface risk, and propose a phased plan.
Book a 30-min architecture review