Web Engineering
Web applications built for speed, accessibility and SEO with React, Next.js and TypeScript. Server rendering, Core Web Vitals and structured data from day one.
Software & applied AI
Build. Ship. Scale.
Capabilities
Four areas we cover end to end, from design through deployment.
Backend Architecture
Scalable, secure APIs in Node.js and Python on PostgreSQL. We design the data model, the auth story and the system boundaries before the first endpoint gets written.
Cloud & DevOps
CI/CD pipelines, containers and orchestration on AWS, Docker and Kubernetes — plus self-hosted infrastructure when the data can't leave the building.
Applied AI
Language models integrated into real products: assistants, tool-using agents, RAG over your own documents and Spanish voice processing. Via APIs or private models on your own infrastructure.
AI
AI that reaches production
The demo is the easy part. Making it work every day, on real data, at a cost that adds up — that's where we work.
Assistants and agents
Agents that use real tools — query your database, book appointments, run calculations, write back to your systems — not chatbots that only talk. Wired up through MCP and function calling.
RAG over your documents
Semantic search across contracts, catalogues, manuals or support history, with verifiable citations back to the source so nobody has to take the model's word for it.
Spanish voice
Voice-note transcription and natural audio replies, running Whisper and TTS locally. Built for WhatsApp and customer support in LatAm, where voice notes are the dominant format.
Private on-premise models
When data can't leave your infrastructure, we deploy open models on your own hardware. No third-party calls, no per-token bill.
Process automation
Extraction, classification and routing of documents, emails and forms — the repetitive work that eats your team's hours today.
Cost and quality control
Prompt caching, per-task model selection and automated evaluations. We measure cost per completed task, not per request.
Our Work
Products and platforms we've designed, built, and shipped to production.
Renovar Solar
Renovar Soluciones Integrales SolaresCorporate website for a Colombian solar energy company that designs, installs, and maintains photovoltaic systems for homes, businesses, and industry.
Search-optimized corporate presence with structured data and per-service landing pages.
Client project · 2025Next.jsSEOWebgarage-ui
DevsClusterAdmin dashboard and S3 browser for Garage clusters and any S3-compatible storage — AWS S3, MinIO, Cloudflare R2, Backblaze B2 and Wasabi.
Admin credentials stay server-side in an AES-256-GCM encrypted session; uploads stream browser-to-storage through presigned URLs.
In-house product · 2026Next.jsS3InfraNotion Widgets
DevsClusterSecure embeddable widgets for Notion: connect a database, set a PIN, and paste an opaque link into an embed block.
Notion tokens are encrypted at rest and never appear in widget URLs; access is PIN-protected, rate-limited server-side, and revocable in one click.
In-house product · 2026Next.jsNotion APISecurity
Our Tech Stack
Tools we use in production every day, not on slides.
- React
- Next.js
- TypeScript
- Node.js
- Python
- PostgreSQL
- AWS
- Docker
- Kubernetes
- LLMs
Why Choose Us
We're committed to delivering excellence at every step
Fast Delivery
Agile development cycles ensure your product reaches market quickly without sacrificing quality.
Enterprise Quality
Battle-tested code with comprehensive testing, security audits, and performance optimization.
Transparent Process
Regular updates, clear communication, and full visibility into project progress.
Long-term Support
Ongoing maintenance, updates, and technical support to keep your systems running smoothly.
Frequently asked questions
What people ask us most before starting.
What exactly does DevsCluster do?
DevsCluster is a software development consultancy based in Colombia. We build web applications with React and Next.js, APIs and backends in Node.js and Python, cloud infrastructure and DevOps, and applied artificial intelligence solutions. We work end to end: design, development, deployment and maintenance.
What kind of AI solutions do you build?
We integrate language models into real products: assistants and agents that use tools connected to your systems, RAG (semantic search over your own documents with verifiable citations), Spanish voice processing with transcription and synthesis, and document process automation. We work with both commercial APIs and open models deployed on your own infrastructure.
Can you run AI models without sending data to third parties?
Yes. We deploy open models on your own infrastructure, on-premise or in your private cloud. This is the right option when confidentiality, data residency or compliance requirements rule out external APIs. There are no calls to outside providers and no per-token billing.
Which technologies do you specialise in?
React, Next.js and TypeScript on the frontend; Node.js, Python and PostgreSQL on the backend; Docker, Kubernetes and AWS for infrastructure. For AI we work with the major language model APIs, MCP for connecting tools, and self-hosted open models.
Do you work with clients outside Colombia?
Yes. We work remotely with clients across Latin America, Spain and the United States, in both Spanish and English.
How does a project start?
With a conversation to understand the problem, the timeline and the real constraints. From there we propose a concrete scope with deliverables and an estimate. You can reach us through the form on this page or at [email protected].
Ready to Ship?
Tell us what you need to build and we'll come back with a concrete scope.
Or email us directly
[email protected]